Retell AI vs LiveKit Agents

Agent ops

1. ai-native userMy coding agent can provision a complete voice agent end to end — create the agent, attach a number, and place a call — through the API, CLI, or MCP without touching the dashboard

   weight 3 · round to Retell AI
   Retell AI~partialprobed6/10

   Retell ships an official Node/Python SDK with full voice endpoint coverage, a CLI for managing agents and phone numbers, and an MCP server whose meta-tools (list/get/invoke_api_endpoint) expose the entire API — including agent create/update/publish — to MCP clients like Cursor or Claude Code, and a keyless runtime probe confirms the MCP handshake works end-to-end. However, the flagship quick-start walkthrough is dashboard-centric (create in UI, assign number in 'configuration settings', test/call via dashboard button), and no evidence explicitly shows a CLI/API/MCP call sequence that attaches a number and places a live call without touching the UI. Missing for 10: an explicit end-to-end CLI/MCP example showing number-attach and call-placement commands, and independent confirmation that non-dashboard number provisioning/outbound calling works in practice.

   • [claimed-docs] “Official Retell SDKs for Node.js and Python. Typed clients with API key auth, structured errors, and full voice and chat endpoint coverage.”
   • [claimed-docs] “Install the Retell CLI to manage agents, phone numbers, knowledge bases, and other Retell resources from your terminal with simple commands.”
   • [claimed-docs] “Use Retell's MCP server to build and manage voice agents from MCP-capable clients like Cursor, Claude Desktop, and Claude Code via Retell AP…”
   • [claimed-docs] “Retell supports the Model Context Protocol (MCP) so you can build Retell AI voice agents directly from MCP-capable clients (Cursor, Claude D…”
   • [claimed-docs] “Agents: create, update, publish, list, and fetch agent versions.”
   • [probe] “PROBE runtime (recorded 2026-09-05): `npx -y @retell-ai/retell-cli --version` printed `retell 0.13.0 (OpenAPI 3.0.0, catalog v4)` keylessly …”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”
   • [claimed-docs] “Assign your agent to the number in the configuration settings”

   proof ↗⚑ flag

   LiveKit Agents~partialprobed5/10

   LiveKit provides a real CLI (`lk`), a documented telephony/SIP stack for provisioning inbound/outbound trunks and placing calls, and framework docs explicitly note it's 'built for coding agents like Claude Code, Cursor, and Codex,' suggesting API/CLI-driven workflows are possible without the dashboard. However, the MCP support documented is for the agent's own tool-calling (consuming MCP servers as tools), not an MCP server exposing agent-provisioning/number-attachment/call-placing actions to a coding agent, and no single evidenced workflow shows an agent programmatically creating an agent, attaching a number, and placing a call end-to-end via CLI/API in one flow — the Agent Builder path shown is dashboard/browser-based (no-code), which contradicts the 'without touching dashboard' framing. Missing for 10: an MCP server (or CLI/API recipe) that lets a coding agent itself create an agent, provision/attach a phone number, and place a call in one documented end-to-end sequence.

   • [github] “Works seamlessly with LiveKit's telephony stack, allowing your agent to make calls to or receive calls from phones.”
   • [claimed-docs] “LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls.”
   • [claimed-docs] “Outbound trunks are used to place outgoing calls.”
   • [probe] “official CLI documented at https://github.com/livekit/livekit-cli”
   • [probe] “PROBE runtime (recorded 2026-09-05): `lk --version` printed `lk version 2.18.6` after a plain `brew install livekit-cli` — the official CLI …”
   • [claimed-docs] “LiveKit is built for coding agents like Claude Code, Cursor, and Codex.”
   • [github] “MCP support: Native support for MCP. Integrate tools provided by MCP servers with one line of code.”
   • [claimed-docs] “LiveKit Agents has first-class support for Model Context Protocol (MCP) servers.”

   proof ↗⚑ flag

2. ai-native userThe platform's own AI helps me author agents — generating or improving prompts, flows, and test cases from a description

   weight 1 · round to Retell AI
   Retell AI~partialclaimed5/10

   Retell docs mention a 'Generate from prompt' feature where Conductor drafts an agent from a plain-English description (retell-docs-30), directly supporting AI-assisted agent authoring. However, there's no evidence of AI-driven improvement of existing prompts/flows or automatic test-case generation, and no independent/hands-on confirmation of Conductor's output quality. Missing for 10: documentation on iterative prompt/flow refinement by the platform's AI, automated test-case generation, and community or hands-on verification of Conductor's generated agents.

   • [claimed-docs] “click **Generate from prompt** (marked Suggested) to let [Conductor](/conductor/create-agent) draft an agent from a plain-English descriptio…”
   • [claimed-docs] “Retell agents are node-based flows or single prompts, with call handling, a knowledge base, and integrations built in.”
   • [claimed-docs] “Drag-and-drop, node-based flows for structured, high-stakes calls.”
   • [claimed-docs] “Single prompt agent... Define your whole agent with one prompt.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   Evidence shows a no-code 'Agent Builder' for browser prototyping and a testing framework, but nothing indicates the platform itself uses AI to generate or improve prompts, flows, or test cases from a natural-language description of the desired agent.

   • [claimed-docs] “LiveKit Agent Builder to prototype and deploy agents directly in your browser without writing code”
   • [claimed-docs] “LiveKit Agent Builder: Prototype and deploy voice agents directly in your browser, without writing any code.”
   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes. They integrate with your existing test suite using pytest (Python) or V…”
   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes... Agent simulations run end-to-end conversations between your agent and…”

   proof ↗⚑ flag

Build

1. developerBuild a working phone voice agent — prompt, voice, and phone number — and take my first live call within an hour

   weight 3 · round to Retell AI
   Retell AI✓fullcommunity8/10

   Docs walk through the exact flow described: create account, pick/generate a template, set prompt and voice, test in dashboard, assign a phone number, and make a live call, with a 15-minute quickstart and free trial credits removing payment friction. Community feedback confirms the demo/agent works end-to-end for real calls, though it also shows occasional conversational glitches (contradictions, confusion) that are quality issues rather than build-flow blockers. missing for 10: independent third-party timing confirmation that a full build-to-first-call truly completes within an hour, and evidence of phone-number provisioning speed in practice.

   • [claimed-docs] “Build your first Retell AI phone agent in 15 minutes: create an account, pick a template, test in the dashboard, deploy to a phone number, a…”
   • [claimed-docs] “Click the "Test" button to start a web call with your agent... This step is free and doesn't need a phone number or payment method.”
   • [claimed-docs] “click **Generate from prompt** (marked Suggested) to let [Conductor](/conductor/create-agent) draft an agent from a plain-English descriptio…”
   • [claimed-docs] “Click the "Test" button to start a web call with your agent”
   • [claimed-docs] “Assign your agent to the number in the configuration settings”
   • [claimed-docs] “New accounts start with \$10 in free trial credits, no payment method required.”
   • [claimed-docs] “Add custom voices to your Retell agent — search ElevenLabs community voices, import a voice clone, or train a clone for a unique brand-speci…”
   • [community] “Curious what model the dentist bot is running on? Tried it out, was surprisingly good, though eventually it contradicted itself (booked a sl…”
   • [community] “I tried to fool the Dental Office demo trying to book on Sunday or outside of the slots it had indicated, and it did a better job than many …”

   proof ↗⚑ flag

   LiveKit Agents✓fullprobed7/10

   Docs show a <10‑minute voice-assistant quickstart (prompt+voice) plus a dedicated telephony/SIP stack for inbound/outbound calls (trunks, phone number), and runtime probes confirm the framework and self-hosted server actually install and boot without extra keys. Together these cover prompt, voice, and phone number needed for a first live call within an hour. Missing for 10: a single unified, hands-on-verified tutorial that walks through phone-number provisioning and first live call end-to-end (currently voice-quickstart and telephony docs are separate), and independent (non-vendor) confirmation of the 'within an hour' timeline.

   • [claimed-docs] “Build and deploy a simple voice assistant in less than 10 minutes.”
   • [claimed-docs] “Build and deploy a simple voice assistant with Python or Node.js in less than 10 minutes.”
   • [claimed-docs] “LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls.”
   • [claimed-docs] “Enable your voice agent to make or take phone calls.”
   • [claimed-docs] “Outbound trunks are used to place outgoing calls.”
   • [github] “Works seamlessly with LiveKit's telephony stack, allowing your agent to make calls to or receive calls from phones.”
   • [probe] “PROBE runtime (recorded 2026-09-05): `lk --version` printed `lk version 2.18.6` after a plain `brew install livekit-cli` — the official CLI …”
   • [probe] “PROBE runtime (recorded 2026-09-05): `livekit-server --dev` (brew-installed, Apache-2.0 OSS) booted on this machine with NO account or key —…”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”

   proof ↗⚑ flag

2. developerRun conversations in multiple languages, including detecting and switching language mid-call

   weight 2 · round drawn
   Retell AI—none0/10

   No documentation in the evidence pack describes multi-language conversation support, language detection, or mid-call language switching; the only related evidence is a founder comment stating multilingual support is merely 'on our roadmap' rather than shipped, indicating the capability does not currently exist.

   • [community] “Founder said multilingual support is on roadmap: 'It's definitely in our roadmap. After the core product—the voice AI part—becomes humanlike…”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   No evidence in the pack addresses multilingual conversation support, language detection, or dynamic language switching mid-call; the docs cover turn detection, interruption handling, tool use, MCP, and telephony but never mention STT/TTS language selection or switching logic. This is a fair axis for a voice-agent framework (many STT/TTS providers support multi-language), so absence of evidence means 'none' rather than 'na'.

   ⚑ flag

3. founderDesign multi-step conversation flows in a visual builder with branching, states, and handoffs without writing code

   weight 2 · round to Retell AI
   Retell AI✓fullclaimed7/10

   Retell's docs describe node-based, drag-and-drop conversation flows for 'structured, high-stakes calls' as an alternative to single-prompt agents, with function calling enabling transfers, call-ending, and API calls that serve as handoff points — directly matching the branching/states/handoff story without requiring code. The quick-start flow (pick template, test in dashboard, deploy) reinforces a no-code workflow. Missing for 10: independent/hands-on evidence of the visual builder's branching UI itself (community evidence only covers conversational behavior, not the builder), and no detail on how 'states' are represented/connected beyond the general node-based description.

   • [claimed-docs] “Retell agents are node-based flows or single prompts, with call handling, a knowledge base, and integrations built in.”
   • [claimed-docs] “Drag-and-drop, node-based flows for structured, high-stakes calls.”
   • [claimed-docs] “Single prompt agent... Define your whole agent with one prompt.”
   • [claimed-docs] “Function calling lets Retell single or multi-prompt agents take real actions — transfer calls, end calls, book appointments, send SMS, and c…”
   • [claimed-docs] “Function calling transforms your AI agent from a conversational interface into an action-oriented assistant.”
   • [claimed-docs] “Build your first Retell AI phone agent in 15 minutes: create an account, pick a template, test in the dashboard, deploy to a phone number, a…”
   • [claimed-docs] “click **Generate from prompt** (marked Suggested) to let [Conductor](/conductor/create-agent) draft an agent from a plain-English descriptio…”

   proof ↗⚑ flag

   LiveKit Agents~partialclaimed3/10

   LiveKit mentions a browser-based 'Agent Builder' for no-code prototyping of voice agents, but the evidence never describes visual branching, explicit states, or handoff design — the core framework is fundamentally code-first (Python/Node.js) with tools, MCP, and turn-detection logic. missing for 10: documentation of branching/state UI, handoff modeling, or any screenshots/examples of the Agent Builder's flow-design capabilities beyond a vague no-code prototyping claim.

   • [claimed-docs] “Prototype and deploy voice agents directly in your browser, without writing any code.”
   • [claimed-docs] “LiveKit Agent Builder to prototype and deploy agents directly in your browser without writing code”
   • [claimed-docs] “LiveKit Agent Builder: Prototype and deploy voice agents directly in your browser, without writing any code.”

   proof ↗⚑ flag

Personalization

1. developerInject dynamic variables and per-caller context at call time so each conversation is personalized

   weight 2 · round drawn
   Retell AI—none0/10

   The evidence pack covers function calling, webhooks, custom telephony, knowledge base retrieval, CRM field sync, and SDK/CLI/MCP tooling, but nowhere describes injecting dynamic variables or per-caller context (e.g., a startCall/create-call parameter for passing caller-specific data into the prompt/LLM at runtime). This is a fair, plausible capability for a voice-agent platform, so absence of evidence yields 'none' rather than 'na'.

   ⚑ flag

   LiveKit Agents—none0/10

   The evidence pack covers tool use, MCP, telephony, testing, and turn detection, but nothing addresses injecting dynamic variables or per-caller context (e.g., participant/room metadata, job context) into an agent's prompt or session at call start. Missing for 10: any documentation of job/participant metadata APIs, per-call context injection, or dynamic prompt personalization.

   ⚑ flag

2. developerGround the agent on my documents with a built-in knowledge base or RAG so it answers from my content

   weight 2 · round to Retell AI
   Retell AI~partialclaimed6/10

   Retell explicitly ships a knowledge base feature where you can crawl websites or upload documents for the agent to retrieve from, and AI QA scores 'knowledge base accuracy' as a call metric, confirming grounding is a first-class capability. However, evidence is thin — only brief homepage-level mentions rather than a dedicated deep-dive doc on KB architecture/RAG internals, and there's no independent/hands-on confirmation of retrieval quality. Missing for 10: dedicated knowledge-base documentation page detailing chunking/retrieval mechanics, and independent verification that answers are accurately grounded in uploaded content.

   • [claimed-docs] “Retell agents are node-based flows or single prompts, with call handling, a knowledge base, and integrations built in.”
   • [claimed-docs] “Crawl a website or upload documents your agent retrieves from.”
   • [claimed-docs] “AI QA scores Retell calls on hallucination, knowledge base accuracy, latency, sentiment, and tool usage to surface quality trends and issues…”

   proof ↗⚑ flag

   LiveKit Agents~partialclaimed4/10

   LiveKit Agents supports calling external APIs/tools that a developer can use to implement RAG ("Call external APIs or lookup data for RAG"), and has full LLM tool-use and MCP integration for wiring in retrieval systems, but there is no built-in knowledge base, document ingestion, or vector-store/RAG pipeline shipped by the framework itself — developers must bring their own RAG implementation via the tools API. Missing for 10: built-in document indexing/vector store, out-of-the-box knowledge-base feature, first-party RAG pipeline or example showing document grounding end-to-end.

   • [claimed-docs] “Call external APIs or lookup data for RAG.”
   • [claimed-docs] “LiveKit Agents has full support for LLM tool use. This feature allows you to create a custom library of tools to extend your agent's context”
   • [claimed-docs] “LiveKit Agents has full support for LLM tool use.”
   • [github] “MCP support: Native support for MCP. Integrate tools provided by MCP servers with one line of code.”
   • [claimed-docs] “Wrap an MCP server in an \`MCPToolset\` and pass it to the agent's \`tools\` parameter”

   proof ↗⚑ flag

Agent access

1. ai-native userPoint an agent at llms.txt or agent-oriented docs

   weight 2 · round to Retell AI
   Retell AI✓fullprobed9/10

   A direct probe confirms an llms.txt file exists at docs.retellai.com/llms.txt returning HTTP 200 with a structured index of the docs, and Retell's docs are also agent-oriented enough to support MCP-based discovery/management of resources. This directly satisfies the story of pointing an agent at llms.txt or agent-oriented docs. Missing for 10: no independent third-party corroboration of llms.txt usage in the wild.

   • [probe] “PROBE llms.txt: HTTP 200 at https://docs.retellai.com/llms.txt # Retell AI > Explore Retell AI docs to learn how to build, test, deploy, an…”
   • [claimed-docs] “Use Retell's MCP server to build and manage voice agents from MCP-capable clients like Cursor, Claude Desktop, and Claude Code via Retell AP…”

   proof ↗⚑ flag

   LiveKit Agents✓fullprobed8/10

   LiveKit publishes a working llms.txt (verified live at docs.livekit.io/llms.txt, HTTP 200) and agent-friendly markdown doc endpoints (e.g. /agents/.md), plus explicit docs noting the platform is 'built for coding agents like Claude Code, Cursor, and Codex,' confirming intentional support for AI-native doc consumption. missing for 10: no independent third-party confirmation that external agents successfully consume these endpoints in practice, and no dedicated documentation page explaining the llms.txt/agent-doc strategy itself.

   • [probe] “PROBE llms.txt: HTTP 200 at https://docs.livekit.io/llms.txt # LiveKit docs > LiveKit is an open-source platform for building voice, video,…”
   • [probe] “PROBE docs-md: HTTP 200 at https://docs.livekit.io/agents/.md LiveKit docs › Build Agents › Get Started › Introduction --- # Introduction …”
   • [claimed-docs] “LiveKit is built for coding agents like Claude Code, Cursor, and Codex.”

   proof ↗⚑ flag

2. ai-native userRun the product headlessly / in CI for automation

   weight 2 · round to LiveKit Agents
   Retell AI~partialprobed7/10

   Retell ships official Node.js/Python SDKs with API-key auth and a terminal CLI for managing agents, phone numbers, and resources, both of which are scriptable outside the dashboard, and a runtime probe confirms the CLI installs and runs keylessly via npm — all consistent with headless/CI use. However, there is no explicit CI/automation documentation (e.g., GitHub Actions example, testing-in-pipeline guide) confirming an officially supported headless workflow. Missing for 10: explicit CI/pipeline documentation or example, confirmation of non-interactive auth flow for CI secrets, and independent evidence of real-world CI usage.

   • [claimed-docs] “Official Retell SDKs for Node.js and Python. Typed clients with API key auth, structured errors, and full voice and chat endpoint coverage.”
   • [claimed-docs] “Install the Retell CLI to manage agents, phone numbers, knowledge bases, and other Retell resources from your terminal with simple commands.”
   • [claimed-docs] “The Retell CLI lets you manage Retell resources from your terminal.”
   • [claimed-docs] “Retell provides official SDKs for Node.js and Python to simplify integration with our platform.”
   • [probe] “PROBE runtime (recorded 2026-09-05): `npx -y @retell-ai/retell-cli --version` printed `retell 0.13.0 (OpenAPI 3.0.0, catalog v4)` keylessly …”
   • [probe] “official CLI documented at https://docs.retellai.com/get-started/cli”

   proof ↗⚑ flag

   LiveKit Agents✓fullprobed8/10

   LiveKit Agents is a Python/Node.js SDK plus self-hostable server that runs as a headless programmatic participant with no UI dependency; it's pip/npm-installable, has a built-in test framework (pytest/Vitest, agent simulations) suitable for CI, and the server/CLI were verified runtime to work keylessly. missing for 10: no explicit documented CI pipeline example (e.g., GitHub Actions config) or headless-mode confirmation beyond inference from server/testing docs.

   • [claimed-docs] “A programmatic participant is any code that joins a LiveKit room as a participant — this includes AI agents, media processors, or custom log…”
   • [claimed-docs] “The Agents framework isn't limited to AI agents. You can use it to deploy any code that needs to process realtime media and data streams as …”
   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes. They integrate with your existing test suite using pytest (Python) or V…”
   • [claimed-docs] “Agent simulations run end-to-end conversations between your agent and an LLM-driven user, then evaluate the results across the full interact…”
   • [github] “Builtin test framework: Write tests and use judges to ensure your agent is performing as expected.”
   • [probe] “PROBE runtime (recorded 2026-09-05): `livekit-server --dev` (brew-installed, Apache-2.0 OSS) booted on this machine with NO account or key —…”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”

   proof ↗⚑ flag

3. ai-native userPlug MCP servers into this product so it can use their tools

   weight 3 · round drawn
   Retell AI✓fullclaimed8/10

   Retell explicitly documents connecting a live voice/chat agent to a remote MCP server so it can invoke that server's tools during calls, directly matching the story of plugging in external MCP servers for tool use. This is a first-party, well-specified capability (not just Retell exposing its own API as MCP) covering both single- and multi-prompt agents. Missing for 10: independent/hands-on verification that a third-party MCP server's tools work reliably mid-call in production.

   • [claimed-docs] “Connect a Retell single- or multi-prompt agent to a remote MCP server so it can call the server's tools during a live voice or chat conversa…”
   • [claimed-docs] “Connect your single- or multi-prompt agent to a remote Model Context Protocol (MCP) server, and the agent can call that server's tools durin…”
   • [claimed-docs] “Connect your single- or multi-prompt agent to a remote [Model Context Protocol (MCP)] server, and the agent can call that server's tools dur…”
   • [claimed-docs] “Function calling transforms your AI agent from a conversational interface into an action-oriented assistant.”

   proof ↗⚑ flag

   LiveKit Agents✓fullclaimed8/10

   Docs and GitHub explicitly state 'first-class support for Model Context Protocol (MCP) servers' with a documented pattern to wrap an MCP server in an MCPToolset and pass it to the agent's tools parameter, described as one line of code integration. This is first-party documentation without independent hands-on verification of MCP-specific usage. Missing for 10: independent/community confirmation of MCP tool integration working end-to-end, and more detail on multi-server or auth configurations.

   • [claimed-docs] “Wrap an MCP server in an \`MCPToolset\` and pass it to the agent's \`tools\` parameter”
   • [claimed-docs] “Wrap an MCP server in an `MCPToolset` and pass it to the agent's `tools` parameter”
   • [claimed-docs] “LiveKit Agents has first-class support for Model Context Protocol (MCP) servers.”
   • [github] “MCP support: Native support for MCP. Integrate tools provided by MCP servers with one line of code.”
   • [github] “Native support for MCP. Integrate tools provided by MCP servers with one line of code.”

   proof ↗⚑ flag

4. ai-native userConnect an agent via an official MCP server

   weight 3 · round to Retell AI
   Retell AI✓fullprobed9/10

   Retell documents and hosts an official MCP server (mcp.retellai.com) that lets MCP-capable clients like Cursor, Claude Desktop, and Claude Code build and manage Retell voice agents via the Retell API, exposing meta-tools for listing/invoking endpoints. A runtime probe confirms the server completes a full keyless JSON-RPC initialize handshake and exposes its tool set, corroborating the vendor docs. Missing for 10: independent/community usage reports of the MCP server in practice beyond vendor docs and the probe.

   • [claimed-docs] “Use Retell's MCP server to build and manage voice agents from MCP-capable clients like Cursor, Claude Desktop, and Claude Code via Retell AP…”
   • [claimed-docs] “Retell supports the Model Context Protocol (MCP) so you can build Retell AI voice agents directly from MCP-capable clients (Cursor, Claude D…”
   • [claimed-docs] “Retell supports the [Model Context Protocol (MCP)] so you can build Retell AI voice agents directly from MCP-capable clients (Cursor, Claude…”
   • [claimed-docs] “Agents: create, update, publish, list, and fetch agent versions.”
   • [probe] “official MCP server documented at https://docs.retellai.com/get-started/mcp-server”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

5. ai-native userUse an official CLI

   weight 2 · round drawn
   Retell AI✓fullprobed9/10

   Retell ships an official CLI (`@retell-ai/retell-cli`) documented for managing agents, phone numbers, and knowledge bases from the terminal, and a runtime probe confirms it installs and runs keylessly via npx. Missing for 10: independent/community usage reports of the CLI itself (only docs and one probe run corroborate it).

   • [claimed-docs] “Install the Retell CLI to manage agents, phone numbers, knowledge bases, and other Retell resources from your terminal with simple commands.”
   • [claimed-docs] “The Retell CLI lets you manage Retell resources from your terminal.”
   • [probe] “official CLI documented at https://docs.retellai.com/get-started/cli”
   • [probe] “PROBE runtime (recorded 2026-09-05): `npx -y @retell-ai/retell-cli --version` printed `retell 0.13.0 (OpenAPI 3.0.0, catalog v4)` keylessly …”

   proof ↗⚑ flag

   LiveKit Agents✓fullprobed9/10

   LiveKit ships an official CLI (livekit-cli), documented on GitHub and verified hands-on to install and run (`lk --version` works via brew install), confirming it's a real, functional official CLI supporting the agent workflow. missing for 10: no deep documentation of full CLI command surface for agent-specific workflows within the evidence pack.

   • [probe] “official CLI documented at https://github.com/livekit/livekit-cli”
   • [probe] “PROBE runtime (recorded 2026-09-05): `lk --version` printed `lk version 2.18.6` after a plain `brew install livekit-cli` — the official CLI …”

   proof ↗⚑ flag

6. ai-native userDrive the product through a documented public API

   weight 3 · round to Retell AI
   Retell AI✓fullprobed9/10

   Retell documents official Node.js/Python SDKs with typed clients and full endpoint coverage, a CLI for managing resources from the terminal, and API references (e.g., clone-voice endpoint) — and a runtime probe confirms the CLI installs/runs keylessly and the hosted MCP server exposes the entire API via meta-tools, proving the API is genuinely agent-drivable. Missing for 10: a publicly discoverable OpenAPI/swagger spec (all standard paths returned 404 in the probe).

   • [claimed-docs] “Official Retell SDKs for Node.js and Python. Typed clients with API key auth, structured errors, and full voice and chat endpoint coverage.”
   • [claimed-docs] “Install the Retell CLI to manage agents, phone numbers, knowledge bases, and other Retell resources from your terminal with simple commands.”
   • [claimed-docs] “Retell provides official SDKs for Node.js and Python to simplify integration with our platform.”
   • [claimed-docs] “Type safety: Full TypeScript support with autocomplete ... Error handling: Structured error responses with detailed messages”
   • [probe] “official MCP server documented at https://docs.retellai.com/get-started/mcp-server”
   • [probe] “official CLI documented at https://docs.retellai.com/get-started/cli”
   • [probe] “PROBE runtime (recorded 2026-09-05): `npx -y @retell-ai/retell-cli --version` printed `retell 0.13.0 (OpenAPI 3.0.0, catalog v4)` keylessly …”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.retellai.com/openapi.json, https://docs.retellai.com/swagger.json, https://docs.retella…”

   proof ↗⚑ flag

   LiveKit Agents✓fullprobed8/10

   LiveKit Agents ships a well-documented, pip/npm-installable Python and Node.js SDK with extensive public API surface (tools, MCP, turn detection, telephony, testing), plus a llms.txt and hands-on verified installs/runtime confirming the API is real and usable by AI-native developers. Missing for 10: independent third-party API reference/versioning audit and deeper evidence of API stability guarantees beyond docs and probes.

   • [claimed-docs] “The Agents framework lets you add any Python or Node.js program to LiveKit rooms as full realtime participants.”
   • [claimed-docs] “A programmatic participant is any code that joins a LiveKit room as a participant — this includes AI agents, media processors, or custom log…”
   • [claimed-docs] “The Agents framework isn't limited to AI agents. You can use it to deploy any code that needs to process realtime media and data streams as …”
   • [github] “MCP support: Native support for MCP. Integrate tools provided by MCP servers with one line of code.”
   • [claimed-docs] “LiveKit Agents has full support for LLM tool use. This feature allows you to create a custom library of tools to extend your agent's context”
   • [probe] “PROBE llms.txt: HTTP 200 at https://docs.livekit.io/llms.txt # LiveKit docs > LiveKit is an open-source platform for building voice, video,…”
   • [probe] “PROBE docs-md: HTTP 200 at https://docs.livekit.io/agents/.md LiveKit docs › Build Agents › Get Started › Introduction --- # Introduction …”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”

   proof ↗⚑ flag

7. ai-native userIssue scoped/least-privilege API credentials for an agent

   weight 2 · round drawn
   Retell AI—none0/10

   The evidence pack documents API key auth, SDKs, CLI, and an MCP server, but nowhere describes scoped, role-based, or least-privilege API key/credential issuance (e.g., per-agent or per-permission keys) for Retell. This is a fair axis for an API/voice-agent platform, but no docs or probes show scoped credential support, so it is unproven.

   • [claimed-docs] “Official Retell SDKs for Node.js and Python. Typed clients with API key auth, structured errors, and full voice and chat endpoint coverage.”
   • [claimed-docs] “Retell provides official SDKs for Node.js and Python to simplify integration with our platform.”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”
   • [claimed-docs] “Use Retell's MCP server to build and manage voice agents from MCP-capable clients like Cursor, Claude Desktop, and Claude Code via Retell AP…”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The evidence pack covers voice/agent features, tool-calling, MCP, testing, and telephony, but contains no mention of scoped or least-privilege API key/token issuance for agents (e.g., LiveKit's grant-based access tokens or credential scoping). Since LiveKit is a platform with API keys and would plausibly support such scoping, this is an applicable but undocumented axis in the given evidence.

   ⚑ flag

8. ai-native userBuild against official SDKs

   weight 2 · round drawn
   Retell AI✓fullprobed9/10

   Retell documents and ships official typed SDKs for Node.js and Python with API key auth, full endpoint coverage, structured errors, and TypeScript autocomplete, plus an official CLI and MCP server for programmatic/agentic access — all independently confirmed by runtime probes (CLI installs and runs, MCP server completes handshake). Missing for 10: no independent third-party review of SDK quality/DX beyond vendor docs and no public OpenAPI spec discovered.

   • [claimed-docs] “Official Retell SDKs for Node.js and Python. Typed clients with API key auth, structured errors, and full voice and chat endpoint coverage.”
   • [claimed-docs] “Retell provides official SDKs for Node.js and Python to simplify integration with our platform.”
   • [claimed-docs] “Type safety: Full TypeScript support with autocomplete ... Error handling: Structured error responses with detailed messages”
   • [probe] “official CLI documented at https://docs.retellai.com/get-started/cli”
   • [probe] “PROBE runtime (recorded 2026-09-05): `npx -y @retell-ai/retell-cli --version` printed `retell 0.13.0 (OpenAPI 3.0.0, catalog v4)` keylessly …”
   • [probe] “official MCP server documented at https://docs.retellai.com/get-started/mcp-server”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”

   proof ↗⚑ flag

   LiveKit Agents✓fullprobed9/10

   LiveKit provides official Python and Node.js SDKs (livekit-agents), verified hands-on to pip-install and import with no key, plus documented plugin ecosystem, tool use, MCP support, testing frameworks, and telephony integration—clearly an official SDK ecosystem for AI-native building. Missing for 10: independent third-party benchmarking of SDK API stability/versioning beyond community latency complaints unrelated to SDK build story.

   • [claimed-docs] “The Agents framework lets you add any Python or Node.js program to LiveKit rooms as full realtime participants.”
   • [claimed-docs] “LiveKit Agents includes a large ecosystem of open source plugins for a variety of AI providers.”
   • [claimed-docs] “LiveKit Agents has full support for LLM tool use. This feature allows you to create a custom library of tools to extend your agent's context”
   • [github] “MCP support: Native support for MCP. Integrate tools provided by MCP servers with one line of code.”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”
   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes. They integrate with your existing test suite using pytest (Python) or V…”

   proof ↗⚑ flag

9. ai-native userSubscribe to events via webhooks

   weight 2 · round to Retell AI
   Retell AI✓fullclaimed8/10

   Retell has dedicated first-party documentation for webhooks describing real-time event notifications pushed to your application as events occur, enabling event-driven integrations rather than polling — directly matching the story of subscribing to events via webhooks. Missing for 10: no independent/hands-on corroboration of webhook reliability or event catalog completeness beyond the docs themselves.

   • [claimed-docs] “Webhooks allow your application to receive real-time notifications about events that occur in your Retell AI account.”
   • [claimed-docs] “webhooks push data to your application as events happen, making your integrations more efficient and responsive.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   No evidence pack items mention webhooks or an event-subscription mechanism for LiveKit Agents; the docs cover MCP tool integration, telephony, testing, and turn detection, but nothing about webhook-based event notifications.

   ⚑ flag

Agentic features

1. ai-native userGet AI-generated insights and suggestions from my data inside the product

   weight 2 · round to Retell AI
   Retell AI~partialclaimed6/10

   Retell's AI QA feature automatically scores calls on hallucination, knowledge-base accuracy, latency, sentiment, and tool usage to surface quality trends and issues, and custom analytics dashboards let users chart/filter call and chat metrics — both are AI-derived insights from the product's own call data. However, this is scoring/quality analytics rather than proactive generative 'suggestions' (e.g., recommended actions, next-best-response, or coaching tips), and there's no independent/hands-on evidence validating the accuracy or usefulness of these AI-generated insights. Missing for 10: evidence of proactive suggestion/recommendation generation beyond scoring, and third-party corroboration of AI QA insight quality.

   • [claimed-docs] “AI QA scores Retell calls on hallucination, knowledge base accuracy, latency, sentiment, and tool usage to surface quality trends and issues…”
   • [claimed-docs] “AI QA automatically evaluates a sampled set of your calls against rules and metrics you configure.”
   • [claimed-docs] “Build custom Retell analytics dashboards to track call and chat metrics like success rate, latency, cost, and concurrency, with charts, filt…”
   • [claimed-docs] “The Analytics dashboard charts your call and chat data so you can see how your agents are performing over time.”
   • [claimed-docs] “Built-in call success and sentiment scoring, plus custom fields synced to your CRM.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

2. ai-native userSet up automations that run autonomously in the background

   weight 2 · round to LiveKit Agents
   Retell AI~partialcommunity6/10

   Retell agents act autonomously during live calls — function calling lets them independently book appointments, send SMS, transfer calls, and call external APIs without human intervention (retell-docs-5, retell-docs-35), and webhooks push real-time events to downstream systems (retell-docs-12, retell-docs-24). This shows in-call autonomous action-taking, which is a form of background automation once a call is triggered. However, there's no evidence of standalone scheduled/triggered automations running independent of a live voice/chat session (e.g., cron-like outbound campaigns or autonomous multi-step workflows outside conversation context). Missing for 10: evidence of scheduled or event-triggered background automations outside live calls, and independent/hands-on confirmation that autonomous function-calling reliably completes tasks unattended (community reports show some confusion/looping during autonomous task execution, retell-comm-1, retell-comm-3, retell-comm-5).

   • [claimed-docs] “Function calling lets Retell single or multi-prompt agents take real actions — transfer calls, end calls, book appointments, send SMS, and c…”
   • [claimed-docs] “Function calling transforms your AI agent from a conversational interface into an action-oriented assistant.”
   • [claimed-docs] “Webhooks allow your application to receive real-time notifications about events that occur in your Retell AI account.”
   • [claimed-docs] “webhooks push data to your application as events happen, making your integrations more efficient and responsive.”
   • [community] “The AI contradicted itself when layering conditionals - it got confused about morning vs afternoon time, kept asking to repeat despite corre…”
   • [community] “Voice sounds great, but: told it unavailable until next year, it confirmed Feb 4th 'next year' but thought current year was 2022. Also got s…”

   proof ↗⚑ flag

   LiveKit Agents✓fullprobed7/10

   LiveKit Agents workers run as persistent 'programmatic participants' deployed to LiveKit Cloud or self-hosted infrastructure, processing realtime streams and telephony calls autonomously in the background without human intervention (docs-8, docs-20, docs-30, docs-26, gh-2), and this is verified hands-on as a real pip-installable OSS framework with a working self-hosted backbone (probe-rt-2, probe-rt-3). Missing for 10: explicit scheduling/cron-trigger mechanisms, independent third-party reports of long-running unattended stability (the one community mention is about observability tooling gaps, not autonomy failure).

   • [claimed-docs] “A programmatic participant is any code that joins a LiveKit room as a participant — this includes AI agents, media processors, or custom log…”
   • [claimed-docs] “The Agents framework isn't limited to AI agents. You can use it to deploy any code that needs to process realtime media and data streams as …”
   • [claimed-docs] “Run your agent on LiveKit Cloud's global infrastructure.”
   • [claimed-docs] “LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls.”
   • [github] “Works seamlessly with LiveKit's telephony stack, allowing your agent to make calls to or receive calls from phones.”
   • [probe] “PROBE runtime (recorded 2026-09-05): `livekit-server --dev` (brew-installed, Apache-2.0 OSS) booted on this machine with NO account or key —…”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”

   proof ↗⚑ flag

3. ai-native userDelegate tasks to a built-in AI assistant inside the product

   weight 3 · round to Retell AI
   Retell AI~partialclaimed3/10

   Retell's dashboard includes 'Conductor', a built-in AI assistant that can draft an entire agent from a plain-English prompt (docs-30), which is a genuine instance of delegating a task to an in-product AI assistant. However this is a single thin mention with no further detail on scope, limits, or other delegable tasks beyond initial agent creation. Missing for 10: documentation of Conductor's full capabilities/limits, evidence of delegating other tasks (not just agent drafting) to a built-in assistant, and independent/hands-on corroboration of Conductor actually working.

   • [claimed-docs] “click **Generate from prompt** (marked Suggested) to let [Conductor](/conductor/create-agent) draft an agent from a plain-English descriptio…”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

4. ai-native userOperate the product with natural-language commands

   weight 2 · round to Retell AI
   Retell AI✓fullprobed7/10

   Retell supports natural-language operation via Conductor's 'Generate from prompt' feature that drafts an agent from a plain-English description, and via its official MCP server which exposes the full API as agent-callable tools so AI-native clients (Cursor, Claude Desktop, Claude Code) can build/manage agents through natural-language MCP tool calls — confirmed live by a keyless JSON-RPC probe against the hosted MCP endpoint. missing for 10: independent/hands-on evaluation of Conductor's prompt-to-agent quality, and no evidence of NL commands for other everyday operations (e.g., dashboard chat-based control) beyond agent creation/management.

   • [claimed-docs] “click **Generate from prompt** (marked Suggested) to let [Conductor](/conductor/create-agent) draft an agent from a plain-English descriptio…”
   • [claimed-docs] “Use Retell's MCP server to build and manage voice agents from MCP-capable clients like Cursor, Claude Desktop, and Claude Code via Retell AP…”
   • [claimed-docs] “Retell supports the Model Context Protocol (MCP) so you can build Retell AI voice agents directly from MCP-capable clients (Cursor, Claude D…”
   • [claimed-docs] “Retell supports the [Model Context Protocol (MCP)] so you can build Retell AI voice agents directly from MCP-capable clients (Cursor, Claude…”
   • [claimed-docs] “Agents: create, update, publish, list, and fetch agent versions.”
   • [probe] “official MCP server documented at https://docs.retellai.com/get-started/mcp-server”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”

   proof ↗⚑ flag

   LiveKit Agents~partialprobed4/10

   LiveKit ships an 'Agent Builder' that lets users prototype/deploy voice agents in-browser without code, and docs explicitly note the framework is 'built for coding agents like Claude Code, Cursor, and Codex' plus agent-readable llms.txt docs — suggesting some AI-native/no-code operability. However there is no evidence of a native natural-language command interface for operating LiveKit itself (e.g., NL-driven CLI, chatops control plane); the MCP support described is for agents calling external tools, not for controlling LiveKit via NL. Missing for 10: a documented NL-driven interface/CLI for configuring or operating the platform itself, and independent confirmation of the Agent Builder's no-code UX actually being NL-based rather than GUI-based.

   • [claimed-docs] “LiveKit is built for coding agents like Claude Code, Cursor, and Codex.”
   • [claimed-docs] “Prototype and deploy voice agents directly in your browser, without writing any code.”
   • [claimed-docs] “LiveKit Agent Builder to prototype and deploy agents directly in your browser without writing code”
   • [probe] “PROBE llms.txt: HTTP 200 at https://docs.livekit.io/llms.txt # LiveKit docs > LiveKit is an open-source platform for building voice, video,…”
   • [probe] “PROBE docs-md: HTTP 200 at https://docs.livekit.io/agents/.md LiveKit docs › Build Agents › Get Started › Introduction --- # Introduction …”

   proof ↗⚑ flag

Api quality

1. ai-native userExplore an interactive API reference with runnable examples

   weight 2 · round drawn
   Retell AI—none0/10

   Evidence shows Retell has API reference doc pages (e.g., api-references/clone-voice.md) and SDKs/CLI/MCP tooling, but no evidence of an interactive, runnable API reference (e.g., embedded 'try it' console or Swagger UI). The probe explicitly found no OpenAPI/Swagger spec exposed (all candidate paths 404), undermining any claim of an interactive reference.

   • [claimed-docs] “Clone a voice from audio files”
   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.retellai.com/openapi.json, https://docs.retellai.com/swagger.json, https://docs.retella…”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   Evidence shows extensive quickstart guides, markdown-based docs (llms.txt), and a no-code 'Agent Builder' prototyping tool, but nothing describes an interactive API reference (e.g., embedded code sandbox, live runnable examples, or Swagger-like explorer) for exploring the SDK/API itself.

   ⚑ flag

2. ai-native userDownload a machine-readable API spec (OpenAPI or equivalent)

   weight 2 · round drawn
   Retell AI—none0/10

   A direct probe for standard OpenAPI/Swagger spec locations (openapi.json, swagger.json, etc.) on Retell's docs domain returned 404 for all candidates, and no evidence pack item shows a downloadable, machine-readable API spec being published or linked from docs. The CLI's version string mentions an internal 'OpenAPI 3.0.0' schema, but this is not shown to be an artifact users can download or fetch programmatically.

   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.retellai.com/openapi.json, https://docs.retellai.com/swagger.json, https://docs.retella…”
   • [probe] “PROBE runtime (recorded 2026-09-05): `npx -y @retell-ai/retell-cli --version` printed `retell 0.13.0 (OpenAPI 3.0.0, catalog v4)` keylessly …”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The evidence pack documents LiveKit's SDKs, CLI, docs (llms.txt), and framework capabilities but contains no mention of a downloadable OpenAPI/Swagger spec or other machine-readable API description for LiveKit's server or Agents APIs. Since LiveKit exposes a real API surface (server API, cloud API), this axis applies but is unevidenced.

   • [probe] “official CLI documented at https://github.com/livekit/livekit-cli”
   • [probe] “PROBE llms.txt: HTTP 200 at https://docs.livekit.io/llms.txt # LiveKit docs > LiveKit is an open-source platform for building voice, video,…”
   • [probe] “PROBE docs-md: HTTP 200 at https://docs.livekit.io/agents/.md LiveKit docs › Build Agents › Get Started › Introduction --- # Introduction …”

   proof ↗⚑ flag

3. ai-native userTest against a sandbox environment without touching production data

   weight 1 · round to LiveKit Agents
   Retell AI~partialclaimed4/10

   Retell's quick-start lets you click 'Test' to run a free web call with your agent before it ever touches a phone number or payment method, and new accounts get $10 trial credit to experiment risk-free — a lightweight way to try an agent without production telephony traffic. However, there's no documented dedicated sandbox environment, test-mode API flag, or explicit data isolation guarantee separating test calls from production records/analytics. Missing for 10: explicit sandbox/staging environment concept, test vs prod data isolation guarantees, and independent confirmation that test-call data doesn't mix with production analytics/QA.

   • [claimed-docs] “Click the "Test" button to start a web call with your agent... This step is free and doesn't need a phone number or payment method.”
   • [claimed-docs] “Click the "Test" button to start a web call with your agent”
   • [claimed-docs] “New accounts start with \$10 in free trial credits, no payment method required.”

   proof ↗⚑ flag

   LiveKit Agents~partialprobed6/10

   LiveKit provides a self-hostable dev server (`livekit-server --dev`) that boots locally with placeholder keys, separate from any production deployment, plus a built-in testing framework with behavioral tests and LLM-driven agent simulations that evaluate agent behavior without needing real production data. However, there is no explicitly branded 'sandbox environment' or staging/production data-isolation feature documented — the sandbox-like capability is inferred from dev-mode self-hosting and test simulations rather than a dedicated sandbox product feature. Missing for 10: an explicit sandbox/staging environment offering with documented separation from production data, and independent confirmation that test simulations never touch production data stores.

   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes. They integrate with your existing test suite using pytest (Python) or V…”
   • [claimed-docs] “Agent simulations run end-to-end conversations between your agent and an LLM-driven user, then evaluate the results across the full interact…”
   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes... Agent simulations run end-to-end conversations between your agent and…”
   • [github] “Builtin test framework: Write tests and use judges to ensure your agent is performing as expected.”
   • [probe] “PROBE runtime (recorded 2026-09-05): `livekit-server --dev` (brew-installed, Apache-2.0 OSS) booted on this machine with NO account or key —…”

   proof ↗⚑ flag

4. ai-native userRely on versioned APIs with a documented deprecation policy

   weight 2 · round drawn
   Retell AI—none0/10

   The evidence pack covers SDKs, CLI, MCP server, and API endpoints extensively, but there is no documentation of API versioning scheme or a deprecation policy for breaking changes. The probe notes an OpenAPI version string (3.0.0, catalog v4) but this is not evidence of a documented deprecation/versioning policy for API consumers.

   ⚑ flag

   LiveKit Agents—none0/10

   No evidence in the pack mentions API versioning scheme, version compatibility guarantees, or a documented deprecation policy for LiveKit Agents' SDKs or APIs.

   ⚑ flag

1. ai-native userPerform bulk operations across many items at once

   weight 2 · round to Retell AI
   Retell AI~partialprobed4/10

   Retell exposes a full-coverage SDK, CLI, and MCP server that let an AI agent script repeated single-item operations (create/update/list agents, phone numbers, knowledge bases) programmatically, enabling scripted bulk-like loops, but there is no documented native batch/bulk endpoint operating on many items in a single call. missing for 10: explicit batch/bulk API endpoints or CLI commands operating on multiple items in one call, evidence of rate-limit-safe bulk workflows, and hands-on confirmation of bulk usage at scale

   • [claimed-docs] “Install the Retell CLI to manage agents, phone numbers, knowledge bases, and other Retell resources from your terminal with simple commands.”
   • [claimed-docs] “Official Retell SDKs for Node.js and Python. Typed clients with API key auth, structured errors, and full voice and chat endpoint coverage.”
   • [claimed-docs] “Use Retell's MCP server to build and manage voice agents from MCP-capable clients like Cursor, Claude Desktop, and Claude Code via Retell AP…”
   • [probe] “PROBE runtime (recorded 2026-09-05): `npx -y @retell-ai/retell-cli --version` printed `retell 0.13.0 (OpenAPI 3.0.0, catalog v4)` keylessly …”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”
   • [claimed-docs] “Agents: create, update, publish, list, and fetch agent versions.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

2. ai-native userDefine rules that trigger actions automatically on events

   weight 3 · round to Retell AI
   Retell AI~partialclaimed6/10

   Retell supports event-driven automation via webhooks that fire on account events (retell-docs-12, retell-docs-24), function calling that lets agents automatically transfer calls, end calls, book appointments, or call external APIs based on conversation logic (retell-docs-5, retell-docs-35), and node-based flows with conditional branching (retell-docs-25) plus AI QA rules evaluated against configured metrics (retell-docs-43). However, this is scattered across call-flow logic and webhook notifications rather than a unified 'if event then action' rule-definition interface. Missing for 10: a dedicated rules/trigger engine UI for arbitrary account-wide events, documentation of webhook-to-action automation chains, and independent verification that rule-based automation works reliably (community reports note conversational logic errors, e.g. retell-comm-1, retell-comm-3, retell-comm-7).

   • [claimed-docs] “Function calling lets Retell single or multi-prompt agents take real actions — transfer calls, end calls, book appointments, send SMS, and c…”
   • [claimed-docs] “Webhooks allow your application to receive real-time notifications about events that occur in your Retell AI account.”
   • [claimed-docs] “webhooks push data to your application as events happen, making your integrations more efficient and responsive.”
   • [claimed-docs] “Drag-and-drop, node-based flows for structured, high-stakes calls.”
   • [claimed-docs] “Function calling transforms your AI agent from a conversational interface into an action-oriented assistant.”
   • [claimed-docs] “AI QA automatically evaluates a sampled set of your calls against rules and metrics you configure.”

   proof ↗⚑ flag

   LiveKit Agents~partialclaimed5/10

   LiveKit Agents supports event-driven automatic behaviors — tool-calling on LLM decisions, turn-detector triggering response timing, and adaptive interruption handling triggering barge-in logic — all of which are 'events auto-trigger actions' patterns built into the agent runtime. However, the evidence shows these as built-in framework behaviors and developer-coded event handlers rather than a user-facing declarative rule engine (e.g., no evidence of a 'when X happens do Y' config surface, webhook/rule subscription API, or no-code rule builder). Missing for 10: explicit rule-definition/webhook-trigger API exposed to non-developer users, documentation of a generalized event-subscription system beyond tool calls and turn logic, and independent confirmation of custom rule automation in production use.

   • [claimed-docs] “Call external APIs or lookup data for RAG.”
   • [claimed-docs] “LiveKit Agents has full support for LLM tool use. This feature allows you to create a custom library of tools to extend your agent's context”
   • [claimed-docs] “LiveKit Agents has full support for LLM tool use.”
   • [claimed-docs] “LiveKit's `TurnDetector` is an audio model that encodes user audio directly, capturing both _what_ is said and _how_ it's said.”
   • [claimed-docs] “A turn detector model can predict that they have more to say and wait for them to finish before responding.”
   • [claimed-docs] “Adaptive interruption handling allows an agent to respond naturally when users speak mid-response... to identify intentional interruptions (…”
   • [claimed-docs] “the model analyzes the acoustic signals to identify intentional interruptions (barge-ins) from conversational backchanneling”
   • [claimed-docs] “Adaptive interruption handling allows an agent to respond naturally when users speak mid-response.”

   proof ↗⚑ flag

3. ai-native userSchedule recurring jobs or workflows

   weight 2 · round drawn
   Retell AI—none0/10

   Retell's docs cover webhooks, function calling, MCP tool-calling, and telephony integration, but nothing describes scheduling recurring jobs, campaigns, or workflows (e.g., cron-like recurring outbound call batches or repeating automations). This is a fair question for a voice-agent platform (buyers often want scheduled/recurring outbound campaigns), so the axis applies, but no evidence in the pack shows this capability. missing for 10: any mention of a scheduler, recurring campaign/job feature, or cron-style automation trigger.

   ⚑ flag

   LiveKit Agents—none0/10

   The evidence pack covers LiveKit Agents' realtime voice/video agent capabilities, tool use, MCP integration, telephony, and testing, but contains no mention of any scheduling, cron-like recurring job, or workflow-automation trigger mechanism. This is a fair capability to ask about for an automation-focused agent framework, but no evidence supports it.

   ⚑ flag

4. ai-native userVersion, review, and roll back my automations

   weight 1 · round drawn
   Retell AI~partialclaimed4/10

   Retell supports 'agent versions' with create/update/publish/list/fetch operations via CLI and MCP (retell-docs-47), implying some versioning and possibly rollback via republishing an older version, but there is no documented review workflow (diffs, approvals, changelogs) or explicit rollback command/UI. missing for 10: explicit rollback mechanism, version diff/review UI, changelog or audit trail, and any independent/hands-on confirmation that version history and rollback actually work as described.

   • [claimed-docs] “Agents: create, update, publish, list, and fetch agent versions.”
   • [claimed-docs] “Install the Retell CLI to manage agents, phone numbers, knowledge bases, and other Retell resources from your terminal with simple commands.”
   • [claimed-docs] “Use Retell's MCP server to build and manage voice agents from MCP-capable clients like Cursor, Claude Desktop, and Claude Code via Retell AP…”

   proof ↗⚑ flag

   LiveKit Agents~partialclaimed4/10

   The pricing page explicitly claims 'Instant rollback to a previous agent deployment,' and the testing/simulation framework (behavioral tests, LLM-judge simulations) supports a review step before deployment, but there's no documented versioning system, changelog, or diff/history UI for automations. Missing for 10: explicit version history/diffing of agent configs, a documented review/approval workflow beyond test suites, and independent confirmation that rollback works in practice.

   • [claimed-docs] “Instant rollback to a previous agent deployment”
   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes. They integrate with your existing test suite using pytest (Python) or V…”
   • [claimed-docs] “Agent simulations run end-to-end conversations between your agent and an LLM-driven user, then evaluate the results across the full interact…”
   • [github] “Builtin test framework: Write tests and use judges to ensure your agent is performing as expected.”

   proof ↗⚑ flag

Compliance

1. founderMeet call-recording consent and disclosure obligations with per-call recording controls and configurable data retention

   weight 2 · round to Retell AI
   Retell AI~partialclaimed5/10

   Retell documents per-agent configurable data retention that auto-deletes transcripts, recordings, and logs after a set period, directly addressing retention obligations. However, there is no evidence of explicit per-call recording enable/disable controls or built-in consent-disclosure features (e.g., automated recording announcements) that founders would need for two-party consent compliance. Missing for 10: per-call recording on/off toggle, consent-disclosure/announcement mechanism, and any documentation tying these to legal consent requirements.

   • [claimed-docs] “Configure per-agent data retention to automatically delete call and chat data — transcripts, recordings, and logs — after a set period for c…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent. After the retention period expires, call and chat data associated with tha…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   No evidence in the pack mentions call recording, consent disclosure, or data retention controls anywhere in LiveKit Agents docs, GitHub, or community sources; the pack covers telephony, tool use, MCP, testing, and turn detection but nothing about recording/retention compliance features. missing for 10: recording consent/disclosure mechanisms, per-call recording toggles, configurable data retention policies, and any documentation or hands-on evidence of these compliance controls.

   ⚑ flag

2. platform-engineerRun regulated workloads with HIPAA/BAA support, SOC 2, and data-residency options

   weight 2 · round drawn
   Retell AI—none0/10

   Evidence shows only generic data-retention controls (docs-15/27/44) but no mention of HIPAA, BAA, SOC 2 certification, or data-residency options anywhere in the docs or community pack, despite this being a fair compliance axis for a voice-AI platform handling call recordings/transcripts.

   • [claimed-docs] “Configure per-agent data retention to automatically delete call and chat data — transcripts, recordings, and logs — after a set period for c…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent. After the retention period expires, call and chat data associated with tha…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The evidence pack contains no mention of HIPAA/BAA agreements, SOC 2 certification, or data-residency controls anywhere in LiveKit's docs, GitHub repo, or probes; self-hosting and TURN/security items only cover TLS/SSL and self-hosted deployment, not compliance attestations. As a platform serving enterprise/regulated workloads, this axis clearly applies, so lack of evidence yields 'none'.

   • [claimed-docs] “The good news is LiveKit includes an embedded TURN server. It's a secure TURN implementation that has integrated authentication with the res…”
   • [claimed-docs] “In order to have a secure LiveKit deployment, you will need a domain as well as a SSL certificate for that domain.”

   proof ↗⚑ flag

Scale

1. platform-engineerSee documented concurrency limits and scale to many simultaneous calls without manual capacity begging

   weight 2 · round to Retell AI
   Retell AI~partialclaimed5/10

   Retell documents a concurrency configuration feature (reserved inbound concurrency vs. outbound limit) showing platform-level concurrency controls exist, but there is no documented default/numeric concurrency limits per plan, no described self-service process to raise limits without contacting support, and no evidence (docs or community) confirming scaling to many simultaneous calls happens without manual requests. missing for 10: published default concurrency limits, self-service limit-increase mechanism, and independent confirmation of large-scale concurrent call handling without support intervention.

   • [claimed-docs] “Reserved inbound concurrency protects inbound calls from being crowded out by outbound traffic.”
   • [claimed-docs] “When `reserved_inbound_concurrency` is configured, outbound calls can use at most your concurrency limit minus the reserved amount.”

   proof ↗⚑ flag

   LiveKit Agents~partialcommunity3/10

   Docs mention running agents on LiveKit Cloud's global infrastructure and self-hosting with an embedded TURN server, implying scale-out capability, but no evidence pack item cites concrete documented concurrency limits, per-instance call caps, or autoscaling guarantees; a community report even flags lack of visibility into per-call cost/latency at scale. missing for 10: published concurrency/capacity numbers, autoscaling documentation, load-testing benchmarks, and confirmation that scaling requires no manual quota requests.

   • [claimed-docs] “Run your agent on LiveKit Cloud's global infrastructure.”
   • [claimed-docs] “The good news is LiveKit includes an embedded TURN server. It's a secure TURN implementation that has integrated authentication with the res…”
   • [community] “We've been running multiple voice AI agents on LiveKit and kept running into visibility issues — no way to measure TTFT, latency across STT …”

   proof ↗⚑ flag

Self host

1. platform-engineerSelf-host the voice agent runtime from open-source code on my own infrastructure

   weight 3 · round to LiveKit Agents
   Retell AI—none0/10

   Retell AI is a hosted SaaS voice-agent platform; all evidence points to cloud-hosted APIs, SDKs, CLI, and MCP server that call Retell's own servers, with no mention of open-source runtime code or self-hosting instructions. There is no evidence a platform-engineer can download and run the voice agent runtime on their own infrastructure.

   • [claimed-docs] “Build your first Retell AI phone agent in 15 minutes: create an account, pick a template, test in the dashboard, deploy to a phone number, a…”
   • [claimed-docs] “Retell starts at $0, pay only for what you use.”
   • [probe] “PROBE runtime (recorded 2026-09-05): `npx -y @retell-ai/retell-cli --version` printed `retell 0.13.0 (OpenAPI 3.0.0, catalog v4)` keylessly …”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”

   proof ↗⚑ flag

   LiveKit Agents✓fullprobed9/10

   LiveKit Agents is open source, pip/npm installable and runs the agent runtime independent of any vendor account, and the underlying WebRTC/SIP server (livekit-server) is Apache-2.0 and self-hostable, confirmed via hands-on probes (booting `livekit-server --dev` locally with no keys, and installing livekit-agents/CLI keylessly) plus official self-hosting deployment docs covering TURN, TLS/domain setup. Missing for 10: no first-party production-scale self-hosting case study or independent report of large-scale self-hosted deployment beyond dev-mode probe.

   • [probe] “PROBE runtime (recorded 2026-09-05): `lk --version` printed `lk version 2.18.6` after a plain `brew install livekit-cli` — the official CLI …”
   • [probe] “PROBE runtime (recorded 2026-09-05): `livekit-server --dev` (brew-installed, Apache-2.0 OSS) booted on this machine with NO account or key —…”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”
   • [claimed-docs] “The good news is LiveKit includes an embedded TURN server. It's a secure TURN implementation that has integrated authentication with the res…”
   • [claimed-docs] “In order to have a secure LiveKit deployment, you will need a domain as well as a SSL certificate for that domain.”
   • [claimed-docs] “A programmatic participant is any code that joins a LiveKit room as a participant — this includes AI agents, media processors, or custom log…”

   proof ↗⚑ flag

Latency

1. platform-engineerSee documented end-to-end voice latency numbers or tuning guidance backing the platform's speed claims

   weight 3 · round to Retell AI
   Retell AI~partialclaimed5/10

   Retell docs offer tuning guidance for latency (transcription mode trade-offs, endpointing settings) and let customers monitor latency via analytics dashboards and AI QA scoring, but there are no documented end-to-end latency benchmarks or numbers substantiating speed claims. missing for 10: published end-to-end latency figures/benchmarks, and independent verification of speed claims.

   • [claimed-docs] “Pick the right Retell transcription mode to balance accuracy against latency — interim results for speed or context-aware results for higher…”
   • [claimed-docs] “Real time transcription is often a trade off between latency and accuracy.”
   • [claimed-docs] “optimize for speed: uses the latest interim results with a low endpointing setting... optimize for accuracy: uses the results with a higher …”
   • [claimed-docs] “Build custom Retell analytics dashboards to track call and chat metrics like success rate, latency, cost, and concurrency, with charts, filt…”
   • [claimed-docs] “The Analytics dashboard charts your call and chat data so you can see how your agents are performing over time.”
   • [claimed-docs] “AI QA scores Retell calls on hallucination, knowledge base accuracy, latency, sentiment, and tool usage to surface quality trends and issues…”
   • [claimed-docs] “AI QA automatically evaluates a sampled set of your calls against rules and metrics you configure.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The evidence pack contains no documented end-to-end latency benchmarks or explicit tuning guidance for reducing turn-taking latency — only feature descriptions (turn detector, adaptive interruption handling) without numbers or configuration guidance. A community report explicitly states users could not measure TTFT or STT→LLM→TTS latency, reinforcing that this visibility/documentation is absent.

   • [claimed-docs] “Adaptive interruption handling allows an agent to respond naturally when users speak mid-response... to identify intentional interruptions (…”
   • [claimed-docs] “LiveKit's `TurnDetector` is an audio model that encodes user audio directly, capturing both _what_ is said and _how_ it's said.”
   • [claimed-docs] “A turn detector model can predict that they have more to say and wait for them to finish before responding.”
   • [claimed-docs] “the model analyzes the acoustic signals to identify intentional interruptions (barge-ins) from conversational backchanneling”
   • [community] “We've been running multiple voice AI agents on LiveKit and kept running into visibility issues — no way to measure TTFT, latency across STT …”

   proof ↗⚑ flag

Turn taking

1. developerRely on the agent to handle interruptions (barge-in) gracefully — stopping speech, updating context, and recovering the turn

   weight 3 · round to LiveKit Agents
   Retell AI—none0/10

   The evidence pack covers transcription latency/accuracy trade-offs (retell-docs-9, retell-docs-39, retell-docs-48) but nowhere describes explicit barge-in/interruption handling — stopping agent speech, updating context, and recovering the turn when a caller interrupts. Community threads show turn-taking confusion (contradictions, loops, disconnects) but do not address interruption handling specifically. Missing for full/partial credit: any documentation or hands-on report confirming barge-in detection, speech-stop behavior, or turn recovery mechanics.

   ⚑ flag

   LiveKit Agents✓fullprobed8/10

   LiveKit Agents has well-documented, purpose-built features for barge-in: adaptive interruption handling distinguishes intentional interruptions from backchanneling, and the TurnDetector model analyzes acoustic and semantic signals to manage turn-taking recovery. These are first-party docs with specific technical detail on the mechanism (stopping speech, updating context, recovering turn), and the runtime probes confirm the framework is real and installable. Missing for 10: independent/hands-on evidence specifically testing barge-in behavior in production (only community evidence is about latency observability, not interruption handling itself) and no third-party benchmark of interruption accuracy.

   • [claimed-docs] “Adaptive interruption handling allows an agent to respond naturally when users speak mid-response... to identify intentional interruptions (…”
   • [claimed-docs] “the model analyzes the acoustic signals to identify intentional interruptions (barge-ins) from conversational backchanneling”
   • [claimed-docs] “LiveKit's `TurnDetector` is an audio model that encodes user audio directly, capturing both _what_ is said and _how_ it's said.”
   • [claimed-docs] “A turn detector model can predict that they have more to say and wait for them to finish before responding.”
   • [claimed-docs] “Adaptive interruption handling allows an agent to respond naturally when users speak mid-response.”
   • [claimed-docs] “LiveKit's TurnDetector is an audio model that encodes user audio directly, capturing both what is said and how it's said.”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”

   proof ↗⚑ flag

2. developerEnable noise suppression or audio filtering so the agent stays coherent on noisy real-world calls

   weight 1 · round drawn
   Retell AI—none0/10

   No evidence pack item mentions noise suppression, background-noise filtering, or audio-preprocessing features for handling noisy real-world call environments; the closest related feature (transcription-mode trade-off between latency and accuracy) addresses turn-taking speed, not noise robustness. This is a plausible axis for a voice-agent platform, so absence of evidence yields 'none' rather than 'na'.

   ⚑ flag

   LiveKit Agents—none0/10

   The evidence pack covers turn detection, interruption handling, telephony, MCP tools, and testing, but contains no mention of noise suppression or audio filtering plugins/features anywhere. Missing for 10: any docs or plugin reference to noise cancellation (e.g. Krisp/BVC), background noise filtering, or audio pre-processing configuration.

   ⚑ flag

3. developerUse model-based end-of-turn detection beyond simple VAD silence timeouts so the agent doesn't talk over slow speakers

   weight 2 · round to LiveKit Agents
   Retell AI—none0/10

   The transcription-mode docs only describe tuning an 'endpointing setting' (low vs high) as a speed/accuracy tradeoff for VAD-style silence detection, not a distinct model-based semantic end-of-turn detector that understands incomplete utterances from slow speakers. No evidence pack item describes a dedicated turn-detection model or algorithm beyond adjusting silence-timeout thresholds.

   • [claimed-docs] “Pick the right Retell transcription mode to balance accuracy against latency — interim results for speed or context-aware results for higher…”
   • [claimed-docs] “Real time transcription is often a trade off between latency and accuracy.”
   • [claimed-docs] “optimize for speed: uses the latest interim results with a low endpointing setting... optimize for accuracy: uses the results with a higher …”

   proof ↗⚑ flag

   LiveKit Agents✓fullclaimed9/10

   LiveKit's TurnDetector is a dedicated audio+semantic model (not just VAD silence) that predicts whether a user has finished speaking or has more to say, waiting accordingly, and adaptive interruption handling further distinguishes real barge-ins from backchanneling — directly addressing not talking over slow speakers. This is documented in detail across multiple first-party doc pages plus GitHub feature lists. missing for 10: independent/hands-on benchmark or community validation of the turn-detector's real-world latency/accuracy impact.

   • [claimed-docs] “LiveKit's `TurnDetector` is an audio model that encodes user audio directly, capturing both _what_ is said and _how_ it's said.”
   • [claimed-docs] “A turn detector model can predict that they have more to say and wait for them to finish before responding.”
   • [claimed-docs] “LiveKit's TurnDetector is an audio model that encodes user audio directly, capturing both what is said and how it's said.”
   • [claimed-docs] “Adaptive interruption handling allows an agent to respond naturally when users speak mid-response... to identify intentional interruptions (…”
   • [claimed-docs] “the model analyzes the acoustic signals to identify intentional interruptions (barge-ins) from conversational backchanneling”
   • [claimed-docs] “Adaptive interruption handling allows an agent to respond naturally when users speak mid-response.”

   proof ↗⚑ flag

1. ai-native userDo everything through the API that I can do in the UI

   weight 2 · round to Retell AI
   Retell AI✓fullprobed8/10

   Retell exposes full API coverage via official Node/Python SDKs, a CLI, and a hosted MCP server that dynamically exposes the entire API (list/get/invoke endpoint tools), confirmed by keyless runtime probes actually reaching the CLI and MCP handshake — meaning nearly anything doable in the dashboard (agents, phone numbers, knowledge bases, function calling, voice cloning, analytics) is API/CLI/MCP accessible. missing for 10: no public OpenAPI spec was found (404s across candidate paths), and no independent hands-on confirmation that every single UI feature (e.g., analytics dashboards, AI QA config) has a 1:1 API equivalent.

   • [claimed-docs] “Install the Retell CLI to manage agents, phone numbers, knowledge bases, and other Retell resources from your terminal with simple commands.”
   • [claimed-docs] “Official Retell SDKs for Node.js and Python. Typed clients with API key auth, structured errors, and full voice and chat endpoint coverage.”
   • [claimed-docs] “Use Retell's MCP server to build and manage voice agents from MCP-capable clients like Cursor, Claude Desktop, and Claude Code via Retell AP…”
   • [claimed-docs] “Retell supports the Model Context Protocol (MCP) so you can build Retell AI voice agents directly from MCP-capable clients (Cursor, Claude D…”
   • [probe] “PROBE runtime (recorded 2026-09-05): `npx -y @retell-ai/retell-cli --version` printed `retell 0.13.0 (OpenAPI 3.0.0, catalog v4)` keylessly …”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.retellai.com/openapi.json, https://docs.retellai.com/swagger.json, https://docs.retella…”
   • [claimed-docs] “Agents: create, update, publish, list, and fetch agent versions.”

   proof ↗⚑ flag

   LiveKit Agents~partialprobed6/10

   LiveKit Agents is API/SDK-first (Python/Node.js) with a full programmatic surface for building agents, tools, telephony, and testing, and there is also a no-code 'Agent Builder' UI for prototyping. However, the evidence doesn't confirm that everything achievable in that browser-based Agent Builder UI (or LiveKit Cloud dashboard features like deployment management, rollback) is equally exposed via API/CLI, so parity between UI and API is not fully demonstrated. missing for 10: explicit evidence that Agent Builder's no-code UI actions (and Cloud dashboard deployment/rollback controls) are all reachable via API/CLI, and confirmation of full CLI/API parity with dashboard features.

   • [claimed-docs] “Prototype and deploy voice agents directly in your browser, without writing any code.”
   • [claimed-docs] “LiveKit Agent Builder to prototype and deploy agents directly in your browser without writing code”
   • [claimed-docs] “LiveKit Agent Builder: Prototype and deploy voice agents directly in your browser, without writing any code.”
   • [claimed-docs] “Instant rollback to a previous agent deployment”
   • [probe] “official CLI documented at https://github.com/livekit/livekit-cli”
   • [probe] “PROBE runtime (recorded 2026-09-05): `lk --version` printed `lk version 2.18.6` after a plain `brew install livekit-cli` — the official CLI …”
   • [claimed-docs] “A programmatic participant is any code that joins a LiveKit room as a participant — this includes AI agents, media processors, or custom log…”

   proof ↗⚑ flag

2. ai-native userExport all of my data in open formats and leave

   weight 3 · round drawn
   Retell AI—none0/10

   Docs describe API/SDK/CLI access, data retention (auto-deletion) and analytics dashboards, but nowhere document a data-export or portability feature that would let a user pull all call/chat data, transcripts, and recordings in an open format to leave the platform; retention docs only cover automatic deletion, not export.

   • [claimed-docs] “Configure per-agent data retention to automatically delete call and chat data — transcripts, recordings, and logs — after a set period for c…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent. After the retention period expires, call and chat data associated with tha…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent.”
   • [claimed-docs] “Retell provides official SDKs for Node.js and Python to simplify integration with our platform.”
   • [claimed-docs] “Install the Retell CLI to manage agents, phone numbers, knowledge bases, and other Retell resources from your terminal with simple commands.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The evidence pack documents LiveKit Agents' open-source, self-hostable nature (Apache-2.0 server, pip/npm installable framework) but contains no explicit mention of data export tooling, open data formats for conversation/session logs, or a documented exit/migration path for user data. Self-hosting mitigates lock-in in principle, but that is not the same as a documented 'export all data' capability.

   • [probe] “PROBE runtime (recorded 2026-09-05): `livekit-server --dev` (brew-installed, Apache-2.0 OSS) booted on this machine with NO account or key —…”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”
   • [claimed-docs] “The good news is LiveKit includes an embedded TURN server. It's a secure TURN implementation that has integrated authentication with the res…”

   proof ↗⚑ flag

3. ai-native userRead the product's source under an open license

   weight 2 · round to LiveKit Agents
   Retell AI—none0/10

   Retell AI is a closed, proprietary SaaS platform; the evidence pack shows SDKs, CLI, MCP server and docs but no mention of source code being published under any open license, nor any GitHub repo for the core platform. Missing for 10: any open-source license grant, public source repository, or licensing terms for the core agent/voice engine.

   ⚑ flag

   LiveKit Agents✓fullprobed8/10

   LiveKit Agents is hosted at github.com/livekit/agents and documented as an open-source framework, with hands-on verification that the packages install and run without any account or key (pip/npm) and that the underlying server is Apache-2.0 OSS. The llms.txt docs also explicitly describe LiveKit as 'an open-source platform.' Missing for 10: an explicit license file/badge citation specifically for the livekit/agents repo (only inferred via probe evidence and sibling repo license).

   • [github] “MCP support: Native support for MCP. Integrate tools provided by MCP servers with one line of code.”
   • [probe] “PROBE llms.txt: HTTP 200 at https://docs.livekit.io/llms.txt # LiveKit docs > LiveKit is an open-source platform for building voice, video,…”
   • [probe] “PROBE runtime (recorded 2026-09-05): `livekit-server --dev` (brew-installed, Apache-2.0 OSS) booted on this machine with NO account or key —…”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”
   • [probe] “official CLI documented at https://github.com/livekit/livekit-cli”

   proof ↗⚑ flag

4. ai-native userSelf-host the core product

   weight 3 · round to LiveKit Agents
   Retell AI—none0/10

   Retell AI is presented entirely as a hosted SaaS (cloud dashboard, hosted APIs, hosted MCP server, usage-based pricing) with no mention anywhere in docs of a self-hosted or on-prem deployment option; all evidence points to a fully managed cloud product.

   • [claimed-docs] “Retell starts at $0, pay only for what you use.”
   • [claimed-docs] “Build your first Retell AI phone agent in 15 minutes: create an account, pick a template, test in the dashboard, deploy to a phone number, a…”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”
   • [claimed-docs] “Integrate Retell voice agents with your own telephony provider using elastic SIP trunking or imported numbers from Twilio, Telnyx, and Vonag…”

   proof ↗⚑ flag

   LiveKit Agents✓fullprobed9/10

   LiveKit provides a documented, Apache-2.0 open-source self-hosting deployment guide including embedded TURN server and SSL setup, and this was hands-on verified: `livekit-server --dev` boots with no account or key, and the open-source livekit-agents framework installs and imports keylessly via pip/npm. This directly demonstrates the core product (server + agents framework) can be self-hosted. Missing for 10: independent third-party production self-hosting case study beyond the probe verification.

   • [claimed-docs] “The good news is LiveKit includes an embedded TURN server. It's a secure TURN implementation that has integrated authentication with the res…”
   • [claimed-docs] “In order to have a secure LiveKit deployment, you will need a domain as well as a SSL certificate for that domain.”
   • [probe] “PROBE runtime (recorded 2026-09-05): `livekit-server --dev` (brew-installed, Apache-2.0 OSS) booted on this machine with NO account or key —…”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”
   • [probe] “PROBE runtime (recorded 2026-09-05): `lk --version` printed `lk version 2.18.6` after a plain `brew install livekit-cli` — the official CLI …”

   proof ↗⚑ flag

Pricing

1. founderSee published per-minute or usage pricing and estimate cost per call before committing

   weight 2 · round to Retell AI
   Retell AI✓fullcommunity8/10

   Retell's pricing page explicitly states pay-as-you-go pricing starting at $0 with no annual contracts, offers an interactive calculator to estimate approximate monthly cost by adjusting key parameters, and new accounts get $10 free trial credit to test before paying — directly enabling a founder to estimate call costs before committing. Community discussion also confirms a concrete per-minute rate (17¢/min) is being charged in practice, corroborating usage-based billing exists. missing for 10: the per-minute rate itself is not shown in first-party docs (only inferred from a community comment), and there's no independent verification of the calculator's accuracy.

   • [claimed-docs] “Retell starts at $0, pay only for what you use.”
   • [claimed-docs] “Adjust key parameters to see an approximate monthly cost for your Retell AI voice agents.”
   • [claimed-docs] “Most voice AI platforms require annual contracts before you write a single line of code. Retell starts at $0, pay only for what you use.”
   • [claimed-docs] “New accounts start with \$10 in free trial credits, no payment method required.”
   • [community] “Founder (yanyan_evie) noted voice provider costs differ: 'The different providers have different prices. openai tts & deepgram are cheaper, …”
   • [community] “A commenter noted per-minute pricing (17¢/min) could be exploited: a malicious caller could keep the bot on the line for hours across many p…”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The pack references a pricing URL (livekit.io/pricing) only for an unrelated 'rollback' feature and never quotes actual per-minute or usage rates; a community comment even highlights that operators struggled with 'cost per call' visibility, but no evidence shows published pricing figures a founder could use to estimate costs before committing.

   • [claimed-docs] “Instant rollback to a previous agent deployment”
   • [community] “We've been running multiple voice AI agents on LiveKit and kept running into visibility issues — no way to measure TTFT, latency across STT …”

   proof ↗⚑ flag

1. ai-native userChoose where my data is stored (region/residency)

   weight 2 · round to LiveKit Agents
   Retell AI—none0/10

   No evidence in the pack mentions data residency, region selection, or geographic storage options; the closest related feature is per-agent data retention/deletion (retell-docs-15/27/44), which addresses retention duration, not location of storage. This axis is applicable to a cloud AI voice platform handling call/chat data, so absence of evidence yields 'none'. Missing for 10: any documentation of selectable regions, data residency guarantees, or geo-specific hosting.

   • [claimed-docs] “Configure per-agent data retention to automatically delete call and chat data — transcripts, recordings, and logs — after a set period for c…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent. After the retention period expires, call and chat data associated with tha…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent.”

   proof ↗⚑ flag

   LiveKit Agents~partialprobed5/10

   LiveKit Agents/server is open-source and self-hostable (embedded TURN server, deployment docs, hands-on verified self-host boot), which lets an operator choose their own hosting region/data location, but there is no explicit documented region-selection or data-residency feature for LiveKit Cloud itself. missing for 10: explicit LiveKit Cloud region-selection/data-residency controls, compliance certifications (e.g. GDPR/SOC2 region guarantees), and any customer-facing UI for choosing storage region.

   • [claimed-docs] “The good news is LiveKit includes an embedded TURN server. It's a secure TURN implementation that has integrated authentication with the res…”
   • [claimed-docs] “In order to have a secure LiveKit deployment, you will need a domain as well as a SSL certificate for that domain.”
   • [probe] “PROBE runtime (recorded 2026-09-05): `livekit-server --dev` (brew-installed, Apache-2.0 OSS) booted on this machine with NO account or key —…”

   proof ↗⚑ flag

2. ai-native userPrevent my data from being used to train AI models

   weight 3 · round drawn
   Retell AI—none0/10

   The evidence only covers configurable data retention (auto-deleting transcripts/recordings/logs after a set period) for compliance, not an explicit opt-out or guarantee against using customer data for AI model training. No docs mention training-data usage policy, opt-out toggles, or contractual no-train clauses.

   • [claimed-docs] “Configure per-agent data retention to automatically delete call and chat data — transcripts, recordings, and logs — after a set period for c…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent. After the retention period expires, call and chat data associated with tha…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   No evidence in the pack addresses data usage for AI model training, opt-out policies, or any privacy commitments regarding training data; the evidence pack is entirely about agent framework features (tools, telephony, testing, MCP support). This is a plausible axis for a platform handling user voice/media data, but absent any documentation on training-data usage or opt-out, it cannot be credited as delivered.

   ⚑ flag

3. ai-native userControl data retention and deletion

   weight 2 · round to Retell AI
   Retell AI!disputedcontradicted5/10

   Retell's docs describe per-agent configurable data retention with automatic, permanent deletion of call/chat transcripts, recordings, and logs after a set period (retell-docs-15, retell-docs-27, retell-docs-44), which is solid first-party evidence for the retention-control axis. However, an independent community report describes the company refusing to delete a user's stored data (credit card) on request, forcing the user to escalate to bank disputes/consumer-affairs complaints — a concrete real-world case where deletion did not work as a customer expected (retell-comm-14). Missing for 10: independent verification that call/chat data deletion itself (not just billing data) works as documented, and no public response/resolution to the deletion complaint.

   • [claimed-docs] “Configure per-agent data retention to automatically delete call and chat data — transcripts, recordings, and logs — after a set period for c…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent. After the retention period expires, call and chat data associated with tha…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent.”
   • [community] “Thread titled 'Retellai won't delete my credit card' - a user complained the company would not delete their stored credit card; commenters s…”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The evidence pack contains no mention of data retention policies, recording/session deletion controls, or configurable retention windows for LiveKit Agents. While self-hosting is documented (implying infrastructure control), no explicit retention/deletion feature or API is evidenced.

   ⚑ flag

4. ai-native userOpt out of telemetry and usage tracking

   weight 2 · round drawn
   Retell AI—none0/10

   The evidence pack covers data retention for call/chat data (transcripts, recordings, logs) but contains no mention of telemetry or usage-tracking opt-out controls for the product itself (e.g., CLI/SDK/dashboard analytics collection). Data retention (retell-docs-15/27/44) addresses deletion of customer call data, not opting out of Retell's own telemetry collection.

   ⚑ flag

   LiveKit Agents—none0/10

   No evidence pack item discusses telemetry, usage tracking, analytics collection, or an opt-out mechanism for LiveKit Agents; the framework being open-source and self-hostable is not sufficient evidence of an explicit telemetry opt-out control.

   ⚑ flag

Call control

1. developerEscalate a live call to a human with warm or blind transfer, passing context along

   weight 2 · round to Retell AI
   Retell AI~partialclaimed4/10

   Retell's function-calling docs list 'transfer calls' as a built-in agent action alongside ending calls, booking appointments, etc., which supports the general concept of escalating a call to a human. However, the evidence never distinguishes warm vs. blind transfer modes nor describes passing conversational context/metadata to the receiving human agent. missing for 10: explicit warm-transfer vs blind-transfer configuration, evidence of context/data hand-off to the human agent, and any hands-on confirmation of this feature working in practice.

   • [claimed-docs] “Function calling lets Retell single or multi-prompt agents take real actions — transfer calls, end calls, book appointments, send SMS, and c…”
   • [claimed-docs] “Function calling transforms your AI agent from a conversational interface into an action-oriented assistant.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   LiveKit Agents documents inbound/outbound telephony (SIP trunks, phone calls) but the evidence pack contains no mention of call transfer (warm or blind) or passing conversational context to a human agent during a live call — this specific escalation capability is never described.

   • [claimed-docs] “LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls.”
   • [claimed-docs] “Enable your voice agent to make or take phone calls.”
   • [claimed-docs] “Outbound trunks are used to place outgoing calls.”
   • [github] “Works seamlessly with LiveKit's telephony stack, allowing your agent to make calls to or receive calls from phones.”

   proof ↗⚑ flag

2. developerMy agent can send DTMF keypresses, navigate IVR menus, and detect or leave voicemail

   weight 1 · round drawn
   Retell AI—none0/10

   The evidence pack covers function calling, custom telephony/SIP trunking, transcription, and MCP tool integration, but nowhere mentions DTMF keypress sending, IVR menu navigation, or voicemail detection/leaving capabilities. This is a fair capability question for a telephony voice-agent platform, but no evidence confirms it is supported.

   ⚑ flag

   LiveKit Agents—none0/10

   Evidence confirms LiveKit's telephony/SIP stack supports inbound/outbound calls (docs-26, docs-31, docs-32, gh-2), but nothing in the pack mentions DTMF keypress sending, IVR menu navigation, or voicemail detection/leaving capabilities specifically. Missing for 10: any documentation or example of DTMF tone generation, automated IVR menu traversal, or voicemail detection/answering-machine-detection logic.

   • [claimed-docs] “LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls.”
   • [claimed-docs] “Enable your voice agent to make or take phone calls.”
   • [claimed-docs] “Outbound trunks are used to place outgoing calls.”
   • [github] “Works seamlessly with LiveKit's telephony stack, allowing your agent to make calls to or receive calls from phones.”

   proof ↗⚑ flag

Campaigns

1. founderRun batch outbound call campaigns with scheduling and throughput controls

   weight 2 · round drawn
   Retell AI—none0/10

   Docs describe outbound vs inbound concurrency reservation (retell-docs-11, retell-docs-41) and custom telephony/SIP integration, but there is no evidence of a batch outbound campaign feature — no mention of scheduling calls, uploading contact lists, or campaign-level throughput controls beyond general concurrency limits. Axis is plausible for a voice-agent platform but unsupported by the evidence pack.

   • [claimed-docs] “Reserved inbound concurrency protects inbound calls from being crowded out by outbound traffic.”
   • [claimed-docs] “When `reserved_inbound_concurrency` is configured, outbound calls can use at most your concurrency limit minus the reserved amount.”
   • [claimed-docs] “Integrate Retell voice agents with your own telephony provider using elastic SIP trunking or imported numbers from Twilio, Telnyx, and Vonag…”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   Evidence shows LiveKit Agents can place individual outbound calls via SIP outbound trunks (docs-26, docs-32, docs-31), but there is no evidence of any batch campaign management, call scheduling, or throughput/rate-limiting controls for running outbound calling campaigns at scale — these features are simply not documented anywhere in the pack.

   • [claimed-docs] “LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls.”
   • [claimed-docs] “Outbound trunks are used to place outgoing calls.”
   • [claimed-docs] “Enable your voice agent to make or take phone calls.”
   • [github] “Works seamlessly with LiveKit's telephony stack, allowing your agent to make calls to or receive calls from phones.”

   proof ↗⚑ flag

Numbers

1. developerProvision phone numbers and run both inbound and outbound calls through the platform's API

   weight 3 · round to Retell AI
   Retell AI✓fullprobed8/10

   Retell's docs show phone-number provisioning via native assignment (docs-32) or importing/SIP-trunking your own numbers from Twilio/Telnyx/Vonage (docs-10,40,49), and both inbound and outbound calling are explicitly supported and distinguished (reserved inbound concurrency vs. outbound traffic in docs-11/41). Full API/SDK coverage for voice endpoints (docs-2,33,46) and a CLI to manage phone numbers (docs-3,21) round out programmatic control, with a working keyless CLI/MCP probe corroborating API-level access (retell-probe-rt-1/2). Missing for 10: explicit documentation of a native 'buy a new number' provisioning endpoint (as opposed to importing existing numbers) and a concrete outbound-call code example.

   • [claimed-docs] “Build your first Retell AI phone agent in 15 minutes: create an account, pick a template, test in the dashboard, deploy to a phone number, a…”
   • [claimed-docs] “Integrate Retell voice agents with your own telephony provider using elastic SIP trunking or imported numbers from Twilio, Telnyx, and Vonag…”
   • [claimed-docs] “Reserved inbound concurrency protects inbound calls from being crowded out by outbound traffic.”
   • [claimed-docs] “This guide shows how to integrate Retell agents with your telephony provider and use your own numbers.”
   • [claimed-docs] “When `reserved_inbound_concurrency` is configured, outbound calls can use at most your concurrency limit minus the reserved amount.”
   • [claimed-docs] “You set up a SIP trunk, configure your number to point to it, and import that number to Retell.”
   • [claimed-docs] “Official Retell SDKs for Node.js and Python. Typed clients with API key auth, structured errors, and full voice and chat endpoint coverage.”
   • [claimed-docs] “Retell provides official SDKs for Node.js and Python to simplify integration with our platform.”
   • [claimed-docs] “Install the Retell CLI to manage agents, phone numbers, knowledge bases, and other Retell resources from your terminal with simple commands.”
   • [claimed-docs] “The Retell CLI lets you manage Retell resources from your terminal.”
   • [probe] “PROBE runtime (recorded 2026-09-05): `npx -y @retell-ai/retell-cli --version` printed `retell 0.13.0 (OpenAPI 3.0.0, catalog v4)` keylessly …”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”

   proof ↗⚑ flag

   LiveKit Agents~partialclaimed6/10

   LiveKit's SIP/telephony stack clearly supports inbound and outbound calling ('LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls', 'Outbound trunks are used to place outgoing calls', 'Enable your voice agent to make or take phone calls') and integrates with the agents runtime via SIP trunks. However, the evidence never shows LiveKit itself provisioning phone numbers through its own API — telephony typically requires configuring an external SIP trunk provider, and no docs here describe a native number-provisioning endpoint. Missing for 10: explicit documentation/evidence of a LiveKit API call that provisions/purchases phone numbers directly (rather than just configuring trunks against externally-acquired numbers), and independent/hands-on confirmation of the inbound+outbound call flow.

   • [claimed-docs] “LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls.”
   • [claimed-docs] “Enable your voice agent to make or take phone calls.”
   • [claimed-docs] “Outbound trunks are used to place outgoing calls.”
   • [github] “Works seamlessly with LiveKit's telephony stack, allowing your agent to make calls to or receive calls from phones.”

   proof ↗⚑ flag

Sip

1. platform-engineerConnect my own carrier or PBX via SIP trunking (or import Twilio/Telnyx numbers) instead of being locked to bundled telephony

   weight 2 · round to Retell AI
   Retell AI✓fullclaimed8/10

   Official docs explicitly describe elastic SIP trunking and importing numbers from Twilio, Telnyx, and Vonage, walking through setting up a SIP trunk and pointing/importing numbers into Retell — directly matching the story. This is first-party documentation without independent hands-on corroboration of the SIP flow itself. Missing for 10: independent/community verification of a real SIP trunk setup working end-to-end, and detail on carrier-specific edge cases (codecs, failover, latency).

   • [claimed-docs] “Integrate Retell voice agents with your own telephony provider using elastic SIP trunking or imported numbers from Twilio, Telnyx, and Vonag…”
   • [claimed-docs] “This guide shows how to integrate Retell agents with your telephony provider and use your own numbers.”
   • [claimed-docs] “You set up a SIP trunk, configure your number to point to it, and import that number to Retell.”

   proof ↗⚑ flag

   LiveKit Agents✓fullclaimed7/10

   LiveKit's SIP-based telephony stack lets agents make/receive calls via inbound and outbound SIP trunks, which is the standard mechanism for bringing your own carrier or PBX (docs-26, docs-32, gh-2, docs-31). This is architecture-agnostic SIP, not a bundled/proprietary telephony lock-in, satisfying the platform-engineer's need to connect external trunks. Missing for 10: explicit documented walkthroughs or examples of importing Twilio/Telnyx numbers specifically, and no independent/hands-on confirmation of a real carrier trunk connection succeeding.

   • [claimed-docs] “LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls.”
   • [claimed-docs] “Outbound trunks are used to place outgoing calls.”
   • [github] “Works seamlessly with LiveKit's telephony stack, allowing your agent to make calls to or receive calls from phones.”
   • [claimed-docs] “Enable your voice agent to make or take phone calls.”

   proof ↗⚑ flag

Analytics

1. ai-native userThe platform's AI reviews my calls for me — scoring quality, flagging failures, and analyzing resolution automatically

   weight 2 · round to Retell AI
   Retell AI~partialclaimed6/10

   Retell's AI QA feature automatically scores sampled calls on hallucination, knowledge base accuracy, latency, sentiment, and tool usage, and the Analytics dashboard tracks success rate and other metrics, directly matching the story's core ask of automated call scoring and quality flagging. However, evidence shows AI QA only evaluates a 'sampled set' of calls (not all calls) against user-configured rules, and there's no documented automatic 'resolution analysis' narrative beyond metrics/dashboards, nor independent hands-on validation of AI QA accuracy. missing for 10: evidence of full (non-sampled) coverage, independent/third-party validation of AI QA scoring accuracy, and detail on automated resolution/failure-cause analysis beyond configured metrics.

   • [claimed-docs] “AI QA scores Retell calls on hallucination, knowledge base accuracy, latency, sentiment, and tool usage to surface quality trends and issues…”
   • [claimed-docs] “AI QA automatically evaluates a sampled set of your calls against rules and metrics you configure.”
   • [claimed-docs] “Build custom Retell analytics dashboards to track call and chat metrics like success rate, latency, cost, and concurrency, with charts, filt…”
   • [claimed-docs] “The Analytics dashboard charts your call and chat data so you can see how your agents are performing over time.”
   • [claimed-docs] “Built-in call success and sentiment scoring, plus custom fields synced to your CRM.”

   proof ↗⚑ flag

   LiveKit Agents!disputedcontradicted3/10

   LiveKit ships a dev-time testing framework with 'judges' and agent simulations to evaluate scripted interactions (docs-9/10/25, gh-4), but this is pre-deployment test tooling, not automatic scoring/flagging/resolution-analysis of live production calls. A hands-on community report explicitly says operators running real agents on LiveKit have no built-in way to measure latency or cost per call, let alone automated quality/resolution review, contradicting the idea that the platform reviews calls for you. Missing for 10: production call-level QA scoring, automatic failure flagging on real traffic, resolution/outcome analysis dashboards.

   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes. They integrate with your existing test suite using pytest (Python) or V…”
   • [claimed-docs] “Agent simulations run end-to-end conversations between your agent and an LLM-driven user, then evaluate the results across the full interact…”
   • [github] “Builtin test framework: Write tests and use judges to ensure your agent is performing as expected.”
   • [community] “We've been running multiple voice AI agents on LiveKit and kept running into visibility issues — no way to measure TTFT, latency across STT …”

   proof ↗⚑ flag

2. founderSee call analytics — success rates, durations, outcomes, sentiment — in dashboards without building my own

   weight 2 · round to Retell AI
   Retell AI✓fullclaimed8/10

   Retell docs describe a built-in Analytics dashboard with charts, filters, and breakdowns for success rate, latency, cost, and concurrency, plus AI QA scoring for sentiment, hallucination, and tool usage—directly matching the founder's need for out-of-box call analytics. Missing for 10: independent/hands-on confirmation of dashboard usability and no evidence of exportable reports or deeper outcome breakdowns beyond what's documented.

   • [claimed-docs] “Build custom Retell analytics dashboards to track call and chat metrics like success rate, latency, cost, and concurrency, with charts, filt…”
   • [claimed-docs] “The Analytics dashboard charts your call and chat data so you can see how your agents are performing over time.”
   • [claimed-docs] “AI QA scores Retell calls on hallucination, knowledge base accuracy, latency, sentiment, and tool usage to surface quality trends and issues…”
   • [claimed-docs] “AI QA automatically evaluates a sampled set of your calls against rules and metrics you configure.”
   • [claimed-docs] “Built-in call success and sentiment scoring, plus custom fields synced to your CRM.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The evidence pack shows testing/simulation tools (pytest/Vitest, LLM-driven simulations) but no built-in production analytics dashboard for success rates, durations, outcomes, or sentiment. A community report explicitly confirms this gap — users running LiveKit voice agents said they had 'no way to measure TTFT, latency across STT→LLM→TTS, or even cost per call' and found debugging painful, indicating no such dashboard exists out of the box.

   • [community] “We've been running multiple voice AI agents on LiveKit and kept running into visibility issues — no way to measure TTFT, latency across STT …”
   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes. They integrate with your existing test suite using pytest (Python) or V…”
   • [claimed-docs] “Agent simulations run end-to-end conversations between your agent and an LLM-driven user, then evaluate the results across the full interact…”
   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes... Agent simulations run end-to-end conversations between your agent and…”

   proof ↗⚑ flag

Monitoring

1. platform-engineerMonitor live calls in production and get alerts when agents misbehave or error rates spike

   weight 1 · round to Retell AI
   Retell AI~partialclaimed6/10

   Retell provides real building blocks for production monitoring — a customizable analytics dashboard tracking success rate, latency, cost and concurrency (retell-docs-13/42), AI QA that scores calls for hallucination, sentiment and tool-usage issues (retell-docs-14/43), and webhooks that push real-time event notifications (retell-docs-12/24) which a platform engineer could wire into an alerting pipeline. However there is no documented native alerting/threshold system (e.g., automatic notification when error rates spike or an agent misbehaves) or a live in-call monitoring view — engineers must build that themselves on top of webhooks/dashboard APIs. Missing for 10: built-in threshold-based alerts or anomaly detection, a real-time 'in-progress calls' monitoring view, and any independent evidence that alerting/monitoring works reliably in production.

   • [claimed-docs] “Webhooks allow your application to receive real-time notifications about events that occur in your Retell AI account.”
   • [claimed-docs] “Build custom Retell analytics dashboards to track call and chat metrics like success rate, latency, cost, and concurrency, with charts, filt…”
   • [claimed-docs] “AI QA scores Retell calls on hallucination, knowledge base accuracy, latency, sentiment, and tool usage to surface quality trends and issues…”
   • [claimed-docs] “webhooks push data to your application as events happen, making your integrations more efficient and responsive.”
   • [claimed-docs] “The Analytics dashboard charts your call and chat data so you can see how your agents are performing over time.”
   • [claimed-docs] “AI QA automatically evaluates a sampled set of your calls against rules and metrics you configure.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   No evidence pack items describe production monitoring dashboards, error-rate alerting, or call-quality observability tooling for LiveKit Agents; the only related evidence is a community report explicitly describing the lack of visibility into TTFT, per-call latency, and cost — a gap, not a delivered capability.

   • [community] “We've been running multiple voice AI agents on LiveKit and kept running into visibility issues — no way to measure TTFT, latency across STT …”

   proof ↗⚑ flag

Testing

1. developerTest agents with simulated conversations or evals before putting them on real phone calls

   weight 2 · round to LiveKit Agents
   Retell AI~partialclaimed5/10

   Retell's dashboard has a free 'Test' web-call button that lets developers try an agent before assigning it to a real phone number, and AI QA can score calls on hallucination, accuracy, sentiment, etc., which supports some testing/analytics workflow. However there's no documented feature for automated simulated-conversation test suites or eval scripts run pre-deployment — AI QA appears to operate on sampled real calls rather than synthetic scripted evals. Missing for 10: dedicated simulation/eval framework for scripted test conversations, batch eval tooling, and any independent verification that pre-call testing catches issues before production use.

   • [claimed-docs] “Click the "Test" button to start a web call with your agent... This step is free and doesn't need a phone number or payment method.”
   • [claimed-docs] “Click the "Test" button to start a web call with your agent”
   • [claimed-docs] “AI QA scores Retell calls on hallucination, knowledge base accuracy, latency, sentiment, and tool usage to surface quality trends and issues…”
   • [claimed-docs] “AI QA automatically evaluates a sampled set of your calls against rules and metrics you configure.”
   • [claimed-docs] “Build custom Retell analytics dashboards to track call and chat metrics like success rate, latency, cost, and concurrency, with charts, filt…”

   proof ↗⚑ flag

   LiveKit Agents✓fullclaimed8/10

   Docs explicitly describe both behavioral tests (pytest/Vitest) and agent simulations that run end-to-end conversations with an LLM-driven user, evaluating results before deployment, and GitHub notes a 'builtin test framework' with judges — directly matching the testing-before-real-calls story, complemented by separate telephony/SIP support for real calls. Missing for 10: independent/hands-on validation of the eval/simulation framework and explicit documentation tying test simulations to pre-phone-call validation workflows.

   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes. They integrate with your existing test suite using pytest (Python) or V…”
   • [claimed-docs] “Agent simulations run end-to-end conversations between your agent and an LLM-driven user, then evaluate the results across the full interact…”
   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes... Agent simulations run end-to-end conversations between your agent and…”
   • [claimed-docs] “Behavioral tests verify specific interactions and expected outcomes... using pytest (Python) or Vitest (Node.js)”
   • [github] “Builtin test framework: Write tests and use judges to ensure your agent is performing as expected.”
   • [github] “Works seamlessly with LiveKit's telephony stack, allowing your agent to make calls to or receive calls from phones.”
   • [claimed-docs] “LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls.”

   proof ↗⚑ flag

Post call

1. developerExtract structured data from every call — outcomes, entities, dispositions — delivered via API or webhook after the call

   weight 2 · round to Retell AI
   Retell AI✓fullclaimed7/10

   Retell explicitly supports post-call structured extraction: built-in call success/sentiment scoring plus custom fields synced to CRM, delivered via real-time webhooks and viewable/aggregated in the analytics/AI QA dashboards. This directly covers outcomes (success rate), dispositions (sentiment), and entities (custom fields) delivered via API/webhook as the story requires. Missing for 10: no independent/hands-on confirmation of the specific post-call-analysis JSON schema or webhook payload structure, and no detail on how custom entity fields are defined/configured.

   • [claimed-docs] “Built-in call success and sentiment scoring, plus custom fields synced to your CRM.”
   • [claimed-docs] “Webhooks allow your application to receive real-time notifications about events that occur in your Retell AI account.”
   • [claimed-docs] “webhooks push data to your application as events happen, making your integrations more efficient and responsive.”
   • [claimed-docs] “Build custom Retell analytics dashboards to track call and chat metrics like success rate, latency, cost, and concurrency, with charts, filt…”
   • [claimed-docs] “The Analytics dashboard charts your call and chat data so you can see how your agents are performing over time.”
   • [claimed-docs] “AI QA scores Retell calls on hallucination, knowledge base accuracy, latency, sentiment, and tool usage to surface quality trends and issues…”
   • [claimed-docs] “AI QA automatically evaluates a sampled set of your calls against rules and metrics you configure.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   LiveKit Agents provides tool-calling, MCP integration, and telephony hooks, but nothing in the evidence describes a built-in mechanism for automatically extracting structured call outcomes/entities/dispositions and delivering them via API or webhook after a call ends — that would have to be custom-built by the developer using the tool-calling primitives. Since no such capability is documented, this applicable axis is unmet.

   • [claimed-docs] “Call external APIs or lookup data for RAG.”
   • [claimed-docs] “LiveKit Agents has full support for LLM tool use. This feature allows you to create a custom library of tools to extend your agent's context”
   • [claimed-docs] “LiveKit telephony lets you build AI-powered voice apps that handle inbound and outbound calls.”
   • [claimed-docs] “Outbound trunks are used to place outgoing calls.”

   proof ↗⚑ flag

Tools

1. ai-native userMy voice agent can plug in MCP servers as tool sources so one integration grants it whole toolsets mid-call

   weight 2 · round to LiveKit Agents
   Retell AI✓fullclaimed8/10

   Retell documents connecting single/multi-prompt voice agents to remote MCP servers so the agent can call the server's tools mid-call (retell-docs-6/23/36), directly matching the story, and this is distinct from Retell's own client-side MCP server for building agents. missing for 10: independent third-party corroboration of mid-call MCP tool invocation in production (only first-party docs and a probe of the client-facing MCP server, not the agent-as-MCP-client tool-call path, are available).

   • [claimed-docs] “Connect a Retell single- or multi-prompt agent to a remote MCP server so it can call the server's tools during a live voice or chat conversa…”
   • [claimed-docs] “Connect your single- or multi-prompt agent to a remote Model Context Protocol (MCP) server, and the agent can call that server's tools durin…”
   • [claimed-docs] “Connect your single- or multi-prompt agent to a remote [Model Context Protocol (MCP)] server, and the agent can call that server's tools dur…”
   • [claimed-docs] “Function calling lets Retell single or multi-prompt agents take real actions — transfer calls, end calls, book appointments, send SMS, and c…”
   • [claimed-docs] “Function calling transforms your AI agent from a conversational interface into an action-oriented assistant.”

   proof ↗⚑ flag

   LiveKit Agents✓fullclaimed9/10

   LiveKit Agents has first-class, documented MCP support ("first-class support for Model Context Protocol (MCP) servers", "Integrate tools provided by MCP servers with one line of code") via `MCPToolset`, which wraps an MCP server and passes it directly to the agent's tools parameter — enabling mid-call toolset access for voice agents. This is corroborated across both docs and GitHub README. Missing for 10: independent hands-on validation of actual mid-call MCP toolset usage beyond vendor docs.

   • [claimed-docs] “Wrap an MCP server in an \`MCPToolset\` and pass it to the agent's \`tools\` parameter”
   • [claimed-docs] “Wrap an MCP server in an `MCPToolset` and pass it to the agent's `tools` parameter”
   • [claimed-docs] “LiveKit Agents has first-class support for Model Context Protocol (MCP) servers.”
   • [github] “MCP support: Native support for MCP. Integrate tools provided by MCP servers with one line of code.”
   • [github] “Native support for MCP. Integrate tools provided by MCP servers with one line of code.”

   proof ↗⚑ flag

2. developerMy agent can call external APIs and custom functions mid-conversation and speak the result without awkward dead air

   weight 3 · round to LiveKit Agents
   Retell AI~partialprobed6/10

   Retell's docs clearly document function calling that lets agents call external APIs mid-conversation (retell-docs-5, retell-docs-35) and even connect to remote MCP servers to invoke tools live during a call (retell-docs-6, retell-docs-23, retell-docs-36), confirmed by a keyless runtime probe showing the hosted MCP server exposing invoke_api_endpoint tools (retell-probe-rt-2). However, no evidence specifically addresses the 'no awkward dead air' requirement — there's no documented filler-speech, streaming-response, or latency-masking mechanism specifically tied to function-call execution; transcription-mode docs (retell-docs-9/39/48) discuss general STT latency tradeoffs, not function-call dead-air handling. missing for 10: explicit filler/backchannel speech during function execution, hands-on demonstration of seamless mid-call tool latency, independent confirmation of no dead-air behavior.

   • [claimed-docs] “Function calling lets Retell single or multi-prompt agents take real actions — transfer calls, end calls, book appointments, send SMS, and c…”
   • [claimed-docs] “Function calling transforms your AI agent from a conversational interface into an action-oriented assistant.”
   • [claimed-docs] “Connect a Retell single- or multi-prompt agent to a remote MCP server so it can call the server's tools during a live voice or chat conversa…”
   • [claimed-docs] “Connect your single- or multi-prompt agent to a remote Model Context Protocol (MCP) server, and the agent can call that server's tools durin…”
   • [claimed-docs] “Connect your single- or multi-prompt agent to a remote [Model Context Protocol (MCP)] server, and the agent can call that server's tools dur…”
   • [probe] “PROBE runtime (recorded 2026-09-05): the hosted MCP server at https://mcp.retellai.com completed a full KEYLESS JSON-RPC initialize handshak…”

   proof ↗⚑ flag

   LiveKit Agents✓fullprobed9/10

   LiveKit Agents has full documented support for LLM tool/function calling (custom Python/Node functions calling external APIs mid-conversation) plus MCP server tool integration, combined with adaptive interruption handling and turn-detection to avoid awkward dead air while tools execute. Runtime probes confirm the framework is real and installable, corroborating the docs claims. missing for 10: independent/hands-on evidence specifically showing tool-call latency handled gracefully in a live conversation (only vendor docs cover this exact combination).

   • [claimed-docs] “Call external APIs or lookup data for RAG.”
   • [claimed-docs] “LiveKit Agents has full support for LLM tool use. This feature allows you to create a custom library of tools to extend your agent's context”
   • [claimed-docs] “LiveKit Agents has full support for LLM tool use.”
   • [claimed-docs] “LiveKit Agents has first-class support for Model Context Protocol (MCP) servers.”
   • [claimed-docs] “Wrap an MCP server in an \`MCPToolset\` and pass it to the agent's \`tools\` parameter”
   • [claimed-docs] “Adaptive interruption handling allows an agent to respond naturally when users speak mid-response... to identify intentional interruptions (…”
   • [claimed-docs] “the model analyzes the acoustic signals to identify intentional interruptions (barge-ins) from conversational backchanneling”
   • [probe] “PROBE runtime (recorded 2026-09-05): pypi livekit-agents 1.8.0 installs and `import livekit.agents` succeeds with no key (npm @livekit/agent…”

   proof ↗⚑ flag

Recording

1. platform-engineerRetrieve full call recordings and transcripts programmatically for every call

   weight 2 · round to Retell AI
   Retell AI~partialclaimed6/10

   Docs confirm that Retell stores per-call transcripts and recordings (referenced in data-retention docs) and that full voice/chat API endpoint coverage exists via official SDKs, implying programmatic retrieval, but no evidence pack item explicitly documents a 'get call' or 'list calls' API endpoint returning recording URLs/transcript text, nor examples of pulling them via SDK/CLI. Missing for 10: explicit API reference for call/recording/transcript retrieval endpoints, CLI/SDK code samples showing recording download or transcript fetch, and any independent confirmation of this working end-to-end.

   • [claimed-docs] “Configure per-agent data retention to automatically delete call and chat data — transcripts, recordings, and logs — after a set period for c…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent. After the retention period expires, call and chat data associated with tha…”
   • [claimed-docs] “Retell allows you to configure a data retention period per agent.”
   • [claimed-docs] “Official Retell SDKs for Node.js and Python. Typed clients with API key auth, structured errors, and full voice and chat endpoint coverage.”
   • [claimed-docs] “Retell provides official SDKs for Node.js and Python to simplify integration with our platform.”
   • [claimed-docs] “Webhooks allow your application to receive real-time notifications about events that occur in your Retell AI account.”
   • [claimed-docs] “webhooks push data to your application as events happen, making your integrations more efficient and responsive.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The evidence pack covers agent building, tool use, MCP, telephony, and testing but contains no mention of recording APIs, transcript storage, or programmatic retrieval of call recordings/transcripts for platform engineers.

   ⚑ flag

Transcription

1. developerGet accurate real-time transcription with control over the STT provider, language models, or key terms

   weight 2 · round to LiveKit Agents
   Retell AI~partialcommunity4/10

   Retell docs confirm real-time transcription with a documented latency/accuracy trade-off (interim vs context-aware endpointing), giving developers some control over accuracy tuning, and community notes hint at underlying provider choices (e.g., Deepgram) for voice pipelines. However, there is no documented ability to select or swap STT providers, choose an ASR language model, or configure custom vocabulary/key-term boosting for transcription accuracy. Missing for 10: explicit STT provider selection API, language model choice for transcription, custom vocabulary/key-term boosting support, and independent accuracy benchmarks.

   • [claimed-docs] “Pick the right Retell transcription mode to balance accuracy against latency — interim results for speed or context-aware results for higher…”
   • [claimed-docs] “Real time transcription is often a trade off between latency and accuracy.”
   • [claimed-docs] “optimize for speed: uses the latest interim results with a low endpointing setting... optimize for accuracy: uses the results with a higher …”
   • [community] “Founder (yanyan_evie) noted voice provider costs differ: 'The different providers have different prices. openai tts & deepgram are cheaper, …”

   proof ↗⚑ flag

   LiveKit Agents~partialclaimed6/10

   LiveKit Agents supports swapping STT/LLM providers via its open-source plugin ecosystem and lets you override API key/base URL for any provider (docs-12, docs-24, docs-35), giving real control over transcription and language-model choice. However there is no evidence of support for custom vocabulary/key-term or hotword boosting in STT, and no independent benchmark of transcription accuracy. Missing for 10: key-term/hotword/phrase-hint configuration support, accuracy benchmarks or independent corroboration of transcription quality.

   • [claimed-docs] “LiveKit Agents includes a large ecosystem of open source plugins for a variety of AI providers.”
   • [claimed-docs] “You can use LiveKit Inference to access many of these models directly through LiveKit Cloud, or you can use the open source plugins to conne…”
   • [claimed-docs] “For any provider not included, you can override the API key and base URL at initialization for the LLM, STT, and TTS interfaces in the plugi…”

   proof ↗⚑ flag

Voices

1. founderClone a custom brand voice and use it for my agents, with a documented consent process

   weight 2 · round to Retell AI
   Retell AI~partialclaimed5/10

   Retell documents voice cloning from uploaded audio files (up to 25 files) and lets you attach the cloned voice to an agent via the voice selector, covering the core 'clone and use a custom voice' capability. However, none of the evidence describes any documented consent-verification step (e.g., consent recording, rights attestation) as part of the cloning flow. Missing for 10: an explicit consent-collection/verification mechanism in the clone-voice API or dashboard docs, and any compliance guidance tying voice cloning to consent requirements.

   • [claimed-docs] “Add custom voices to your Retell agent — search ElevenLabs community voices, import a voice clone, or train a clone for a unique brand-speci…”
   • [claimed-docs] “Clone a voice from audio files”
   • [claimed-docs] “Audio files to use for voice cloning. Up to 25 files allowed.”
   • [claimed-docs] “You can also add a voice clone by clicking "Add custom voice" in the voice selector.”

   proof ↗⚑ flag

   LiveKit Agents—none0/10

   The axis applies to this product kind (peer products hold positive or none verdicts on this story), so lack of evidence for an applicable capability is "none", never "na". (na/none harmonized at arena bring-up — see pipeline/scripts/na-harmonize.ts.)

   ⚑ flag

2. developerChoose from a broad voice library or plug in multiple TTS providers to get the voice I want

   weight 2 · round to Retell AI
   Retell AI✓fullcommunity7/10

   Docs confirm a searchable voice library (ElevenLabs community voices) plus voice cloning options (retell-docs-7, retell-docs-37, retell-docs-38), and a founder community comment independently confirms multiple underlying TTS providers (OpenAI TTS, Deepgram, ElevenLabs) with different pricing tiers (retell-comm-10), showing developers can indeed pick across providers/voices. Missing for 10: a first-party docs page enumerating all supported TTS providers and API-level provider-switching parameters beyond the voice-selector UI, and independent hands-on comparison of voice quality across providers.

   • [claimed-docs] “Add custom voices to your Retell agent — search ElevenLabs community voices, import a voice clone, or train a clone for a unique brand-speci…”
   • [claimed-docs] “You can also add a voice clone by clicking "Add custom voice" in the voice selector.”
   • [claimed-docs] “In the voice selector, you can click "Add custom voice" to search and add publicly available community voices.”
   • [community] “Founder (yanyan_evie) noted voice provider costs differ: 'The different providers have different prices. openai tts & deepgram are cheaper, …”

   proof ↗⚑ flag

   LiveKit Agents~partialclaimed6/10

   Docs confirm a large open-source plugin ecosystem for TTS/STT/LLM providers, plus LiveKit Inference for provider access, and an override mechanism for API key/base URL for providers not natively included, supporting multi-provider TTS flexibility. However, there's no explicit mention of a 'voice library' (e.g., curated voice catalog/selection UI) or enumeration of specific TTS providers/voices, so the 'broad voice library' half of the story is unevidenced. Missing for 10: explicit voice catalog/list of supported TTS providers and voices, evidence of ease of switching between TTS voices, independent confirmation of provider breadth.

   • [claimed-docs] “LiveKit Agents includes a large ecosystem of open source plugins for a variety of AI providers.”
   • [claimed-docs] “You can use LiveKit Inference to access many of these models directly through LiveKit Cloud, or you can use the open source plugins to conne…”
   • [claimed-docs] “For any provider not included, you can override the API key and base URL at initialization for the LLM, STT, and TTS interfaces in the plugi…”

   proof ↗⚑ flag