Canva vs Spline

Agent ops

1. ai-native userAn agent can create or modify designs programmatically — through a write-capable API, plugin surface, or MCP tools — not just read them

   weight 2 · round to Canva
   Canva✓fullprobed9/10

   Canva ships an official MCP server explicitly documented to let AI assistants create designs from text, edit existing designs via natural language, manage assets/brand kits, and export/resize outputs — a clear write-capable agent surface, corroborated by a dedicated MCP docs probe. This is backed further by the Connect APIs (write-capable REST API with OpenAPI/SDK generation) and Apps SDK for programmatic design manipulation. missing for 10: independent/hands-on third-party verification that the MCP write operations work reliably in practice.

   • [claimed-docs] “give your users access to Canva's design capabilities through natural language, directly from the AI tools they already use”
   • [claimed-docs] “Design generation: Create designs programmatically from text descriptions and specifications”
   • [claimed-docs] “Design editing: Apply targeted modifications to existing designs through natural language commands”
   • [claimed-docs] “Design discovery: Search and retrieve designs, pages, and folders from your design library”
   • [claimed-docs] “Asset and brand management: Upload and organize assets, manage folders, and access brand kits and templates”
   • [claimed-docs] “Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “You can use the Apps SDK to create apps that import content into Canva, add elements to a user's design, automate common design tasks, and m…”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”

   proof ↗⚑ flag

   Spline✓fullprobed8/10

   Spline ships a dedicated MCP Server (bundled in desktop app) that lets Claude, Cursor, Codex, VS Code, Antigravity and ChatGPT drive the live editor to create/modify designs, plus a write-capable Code API for programmatic scene manipulation and webhooks/Real-time API for external event-driven updates. This goes beyond read-only access into genuine programmatic creation/modification. Missing for 10: independent hands-on validation of the MCP server's write actions and no public OpenAPI spec was found (openapi probe 404s), so full API surface is only documented, not independently verified.

   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [claimed-docs] “connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here — there's nothing to install or configure, the app…”
   • [claimed-docs] “The MCP server is bundled in the desktop app, so connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here —…”
   • [claimed-docs] “the desktop app gives you a few things the browser can't: **tabs** for working across several files, **system fonts**, and **MCP** support”
   • [claimed-docs] “The **Spline MCP Server** connects AI coding tools like **Claude**, **Cursor**, **Codex**, and **Antigravity** directly to your Spline edito…”
   • [claimed-docs] “Spline works fully in the browser, but the desktop app gives you a few things the browser can't: **tabs** for working across several files, …”
   • [claimed-docs] “It enables you to trigger actions, modify object properties, and create dynamic, responsive 3D environments directly from your existing code…”
   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “Webhooks in Spline allow you to create real-time, event-driven interactions by listening for and responding to external events.”
   • [probe] “official MCP server documented at https://docs.spline.design/basics/download-spline-for-desktop”
   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.spline.design/openapi.json, https://docs.spline.design/swagger.json, https://docs.splin…”

   proof ↗⚑ flag

2. ai-native userAn agent can render and export designs — screenshots, images, or production formats — through a documented programmatic surface

   weight 2 · round to Canva
   Canva✓fullprobed8/10

   Canva's Connect APIs and MCP server both provide documented programmatic export capability — the MCP server explicitly supports 'Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, etc.) and resize designs for different channels', and the Connect APIs allow creating/syncing designs and assets with OpenAPI-generated SDKs, confirmed via a live probe of the MCP docs. This directly satisfies an agent's need to render/export designs programmatically. Missing for 10: independent/hands-on confirmation of actual export calls succeeding (no third-party corroboration, only first-party docs), and the OpenAPI spec itself wasn't resolvable via direct probe.

   • [claimed-docs] “Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “Design export:** Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “You can use the Connect APIs to create web app integrations that create and sync assets, designs, comments, and more.”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”

   proof ↗⚑ flag

   Spline~partialprobed4/10

   Spline documents multiple export/programmatic surfaces (export-as-code for Vanilla JS/Three.js/React/Next.js/r3f, self-hosted ZIP export, Code API, Real-time API, webhooks) and an MCP server that lets agents like Claude/Cursor drive the editor, but none of the docs explicitly describe an agent invoking a screenshot/image render or export action through these interfaces — the MCP surface is framed as scene building/editing, not export/rendering. Missing for 10: explicit MCP or API tool for triggering exports/screenshots/renders programmatically, and independent confirmation this works end-to-end for an agent.

   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS, Three.js, React, Next.js, react-three-fiber”
   • [claimed-docs] “The Code API allows you to seamlessly integrate interactive 3D experiences into your projects, offering greater control and customization.”
   • [claimed-docs] “allows you to export your entire project, including all assets and the Spline runtime, as a fully self-contained ZIP file”
   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [claimed-docs] “The MCP server is bundled in the desktop app, so connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here —…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.spline.design/openapi.json, https://docs.spline.design/swagger.json, https://docs.splin…”

   proof ↗⚑ flag

3. ai-native userMy coding agent can pull real design context — frames, variables, screenshots, component mappings — through an MCP server or API to implement the design

   weight 3 · round drawn
   Canva~partialprobed6/10

   Canva has an official documented MCP server enabling AI assistants to search/retrieve designs, pages and folders, manage assets and brand kits, and export in multiple formats (PDF/PNG/JPG/etc.), which covers much of 'pulling real design context' for an agent to act on. However, the story's specific terms — frames, variables, component mappings — aren't explicitly evidenced, and there's no independent/hands-on confirmation that a coding agent can actually reconstruct implementation-level design context via this server. Missing for 10: explicit support for design variables, frame-level structural access, component-to-code mapping, and third-party corroboration of agent workflows.

   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [claimed-docs] “Design discovery: Search and retrieve designs, pages, and folders from your design library”
   • [claimed-docs] “Asset and brand management: Upload and organize assets, manage folders, and access brand kits and templates”
   • [claimed-docs] “Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “Design discovery:** Search and retrieve designs, pages, and folders from your design library”
   • [claimed-docs] “Design export:** Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”

   proof ↗⚑ flag

   Spline~partialprobed6/10

   Spline ships an official MCP server bundled with the desktop app that connects Claude, Cursor, Codex, Antigravity and other MCP clients directly to the open editor, and separately offers a Code API, Real-time API and code export (React/Next.js/Three.js/react-three-fiber) that could let an agent read variables and scene structure to implement designs. However the MCP server's documented purpose is to let AI clients drive/edit the Spline scene ('straight into your open editor tab'), not explicitly to pull frames, screenshots, or component mappings as design-context for an external implementation — that specific dev-mode-style extraction flow is not evidenced. Missing for 10: explicit support for screenshot export via MCP/API, frame-level context retrieval, and documented component-mapping extraction for code implementation.

   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [claimed-docs] “The MCP server is bundled in the desktop app, so connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here —…”
   • [claimed-docs] “The **Spline MCP Server** connects AI coding tools like **Claude**, **Cursor**, **Codex**, and **Antigravity** directly to your Spline edito…”
   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS, Three.js, React, Next.js, react-three-fiber”
   • [claimed-docs] “The Code API allows you to seamlessly integrate interactive 3D experiences into your projects, offering greater control and customization.”
   • [probe] “official MCP server documented at https://docs.spline.design/basics/download-spline-for-desktop”

   proof ↗⚑ flag

Agent access

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

   weight 2 · round to Spline
   Canva✓fullprobed8/10

   Canva confirms an accessible llms.txt file (HTTP 200) plus multiple scoped llms.txt files for sub-docs (e.g., scim/llms.txt, admin/llms.txt, audit-logs/llms.txt) and dedicated .md docs pages (mcp.md, connect.md, apps.md) explicitly structured for agent consumption. missing for 10: independent/third-party confirmation of an agent successfully using llms.txt to navigate docs, and a root-level single canonical llms.txt covering all products rather than fragmented per-section files.

   • [probe] “PROBE llms.txt: HTTP 200 at https://www.canva.dev/llms.txt # Canva Developers Documentation This file lists documentation for Canva product…”
   • [claimed-docs] “Use the SCIM API to automate the provisioning of Canva user accounts for your team.”
   • [claimed-docs] “The Canva Admin API lets you manage teams, groups, and users in a Canva organization.”
   • [claimed-docs] “Canva provides audit logs that contain records of user activities in Canva, such as installing a Canva App, exporting a design for download,…”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [claimed-docs] “give your users access to Canva's design capabilities through natural language, directly from the AI tools they already use”

   proof ↗⚑ flag

   Spline✓fullprobed9/10

   Spline has a verified llms.txt at docs.spline.design/llms.txt (HTTP 200 probe) plus .md-suffixed doc pages that render clean markdown for agent consumption, confirming genuine agent-oriented docs infrastructure. missing for 10: independent third-party confirmation that agents actually consume these docs successfully in practice.

   • [probe] “PROBE llms.txt: HTTP 200 at https://docs.spline.design/llms.txt # Spline Documentation ## Docs - [Getting Started](https://docs.spline.des…”
   • [probe] “PROBE docs-md: HTTP 200 at https://docs.spline.design/basics/what-is-spline.md # What is Spline? (/basics/what-is-spline) Spline is a web…”
   • [claimed-docs] “How Spline compares to traditional 3D tools like Blender, Maya and Cinema 4D.”

   proof ↗⚑ flag

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

   weight 2 · round to Canva
   Canva~partialprobed5/10

   Canva's Connect APIs and bulk/personalization features (canva-docs-2, canva-docs-7, canva-docs-8, canva-docs-16, canva-docs-26) show that programmatic, non-interactive design generation and export are possible, and an OpenAPI description exists for SDK generation (canva-docs-31, canva-docs-36) which could support CI pipelines. However, Connect APIs are OAuth-based integrations meant for user-authorized platforms, not clearly documented as headless/service-account CI automation, and a probe for a hosted OpenAPI spec returned 404s (canva-probe-3), suggesting the automation surface is less mature for pure CI use. Missing for 10: explicit CI/service-account authentication flow, documented headless execution examples, and a verifiable OpenAPI spec endpoint.

   • [claimed-docs] “Use the Canva Connect APIs to integrate Canva capabilities into your platform or your users' workflow.”
   • [claimed-docs] “Automatically personalize and populate Brand Templates with real-time content from your data sources.”
   • [claimed-docs] “Scale design output - generate personalized content in bulk using your own data and visuals.”
   • [claimed-docs] “Automatically create resized versions of designs in new dimensions, formats, or aspect ratios.”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries...to generate client SDKs in vari…”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries... to generate client SDKs in var…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://www.canva.dev/openapi.json, https://www.canva.dev/swagger.json, https://www.canva.dev/api/op…”

   proof ↗⚑ flag

   Spline—none0/10

   Spline is documented purely as a browser/desktop GUI editor with a multiplayer, real-time canvas; its MCP server explicitly requires the desktop app with 'your open editor tab' running interactively, and no CLI, headless render mode, or CI-triggerable build/export process is documented anywhere in the evidence pack. Webhooks and the Real-time API only affect published/runtime scenes, not the design-time editor pipeline, so there is no path to running Spline itself headlessly in automation.

   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [claimed-docs] “The MCP server is bundled in the desktop app, so connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here —…”
   • [claimed-docs] “connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here — there's nothing to install or configure, the app…”
   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “Webhooks in Spline allow you to create real-time, event-driven interactions by listening for and responding to external events.”

   proof ↗⚑ flag

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

   weight 3 · round drawn
   Canva—none0/10

   All evidence describes Canva exposing an official MCP *server* so external AI assistants (Claude, ChatGPT, etc.) can call Canva's own design tools (canva-docs-5, canva-docs-28, canva-probe-4) — the opposite direction of the story, which asks whether Canva itself can act as an MCP client plugging into external MCP servers to use their tools. No evidence shows Canva consuming or configuring external MCP servers.

   • [claimed-docs] “give your users access to Canva's design capabilities through natural language, directly from the AI tools they already use”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [claimed-docs] “If you're a user wanting to connect your personal AI assistant to Canva's AI Connector, see our Help Center”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”

   proof ↗⚑ flag

   Spline—none0/10

   All evidence describes Spline shipping an MCP *server* (docs-13, docs-25, docs-36, docs-53) that external MCP clients like Claude/Cursor/Codex connect to — the reverse of this story, which asks whether a user can plug external MCP servers into Spline so its own AI agent can use their tools. No evidence shows Spline acting as an MCP client or consuming third-party MCP tool servers.

   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [claimed-docs] “The MCP server is bundled in the desktop app, so connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here —…”
   • [claimed-docs] “The **Spline MCP Server** connects AI coding tools like **Claude**, **Cursor**, **Codex**, and **Antigravity** directly to your Spline edito…”
   • [claimed-docs] “The Spline MCP Server connects AI coding tools like Claude, Cursor, Codex, and Antigravity directly to your Spline editor.”
   • [claimed-docs] “connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here — there's nothing to install or configure, the app…”

   proof ↗⚑ flag

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

   weight 3 · round to Canva
   Canva✓fullprobed9/10

   Canva ships an official MCP server (canva-docs-28, canva-probe-4) that lets AI assistants connect via natural language to Canva's design capabilities, with documented features spanning design generation, editing, discovery, export, asset/brand management, and collaboration (canva-docs-17 through 22, 33-35). Missing for 10: independent/hands-on third-party corroboration of the MCP server working in practice.

   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [claimed-docs] “give your users access to Canva's design capabilities through natural language, directly from the AI tools they already use”
   • [claimed-docs] “Design generation: Create designs programmatically from text descriptions and specifications”
   • [claimed-docs] “Design editing: Apply targeted modifications to existing designs through natural language commands”
   • [claimed-docs] “Design discovery: Search and retrieve designs, pages, and folders from your design library”
   • [claimed-docs] “Asset and brand management: Upload and organize assets, manage folders, and access brand kits and templates”
   • [claimed-docs] “Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “Collaborative workflows: Add comments to designs and integrate feedback loops directly into your workflows”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”

   proof ↗⚑ flag

   Spline✓fullprobed8/10

   Spline documents an official MCP server bundled in the desktop app that connects Claude, Cursor, Codex, VS Code, Antigravity, and ChatGPT directly to the editor, requiring no separate install/config. This is a clear first-party MCP server offering with dedicated docs. Missing for 10: independent/hands-on community corroboration of the MCP server actually working, and no detail on the specific tools/capabilities exposed via the server.

   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [claimed-docs] “The MCP server is bundled in the desktop app, so connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here —…”
   • [claimed-docs] “The **Spline MCP Server** connects AI coding tools like **Claude**, **Cursor**, **Codex**, and **Antigravity** directly to your Spline edito…”
   • [claimed-docs] “The Spline MCP Server connects AI coding tools like Claude, Cursor, Codex, and Antigravity directly to your Spline editor.”
   • [probe] “official MCP server documented at https://docs.spline.design/basics/download-spline-for-desktop”

   proof ↗⚑ flag

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

   weight 3 · round to Canva
   Canva✓fullprobed8/10

   Canva provides well-documented Connect APIs with an OpenAPI description for client SDK generation, plus Apps SDK, Admin API, SCIM API, and a first-party MCP server enabling AI assistants to drive Canva programmatically — directly matching an AI-native user's ability to drive the product via public API. Missing for 10: independent hands-on corroboration of API robustness/rate limits and a directly resolvable public OpenAPI file (probe found candidate OpenAPI endpoints 404, though docs claim one exists).

   • [claimed-docs] “Use the Canva Connect APIs to integrate Canva capabilities into your platform or your users' workflow.”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries...to generate client SDKs in vari…”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries... to generate client SDKs in var…”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [claimed-docs] “Design generation: Create designs programmatically from text descriptions and specifications”
   • [claimed-docs] “Design discovery: Search and retrieve designs, pages, and folders from your design library”
   • [claimed-docs] “Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”
   • [probe] “PROBE openapi: all candidate paths 404 (https://www.canva.dev/openapi.json, https://www.canva.dev/swagger.json, https://www.canva.dev/api/op…”

   proof ↗⚑ flag

   Spline~partialprobed7/10

   Spline documents several external-facing interfaces an AI-native user could drive: the Real-time API and Webhooks for event-driven external integration, a Code API for web to control published scenes from external code/UI, and an MCP Server that lets AI coding agents (Claude, Cursor, Codex) manipulate the editor directly. These are genuine documented, public-facing APIs beyond just manual UI use. missing for 10: a formal REST/OpenAPI specification (probe explicitly found no openapi.json/swagger.json at any candidate path), and independent/hands-on corroboration that these APIs work reliably outside vendor docs.

   • [claimed-docs] “It enables you to trigger actions, modify object properties, and create dynamic, responsive 3D environments directly from your existing code…”
   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “Webhooks in Spline allow you to create real-time, event-driven interactions by listening for and responding to external events.”
   • [claimed-docs] “The Code API allows you to seamlessly integrate interactive 3D experiences into your projects, offering greater control and customization.”
   • [claimed-docs] “The **Spline MCP Server** connects AI coding tools like **Claude**, **Cursor**, **Codex**, and **Antigravity** directly to your Spline edito…”
   • [claimed-docs] “The Spline MCP Server connects AI coding tools like Claude, Cursor, Codex, and Antigravity directly to your Spline editor.”
   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.spline.design/openapi.json, https://docs.spline.design/swagger.json, https://docs.splin…”
   • [probe] “official MCP server documented at https://docs.spline.design/basics/download-spline-for-desktop”

   proof ↗⚑ flag

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

   weight 2 · round drawn
   Canva—none0/10

   While Canva exposes Connect APIs, OAuth-based integrations, and an MCP server for AI assistants, none of the evidence describes issuing scoped or least-privilege credentials specifically for an agent (e.g., granular OAuth scopes, per-agent API keys, or restricted permission sets). Private integrations are gated by plan tier (docs-41) rather than granular scoping, and MCP/AI Connector docs discuss connecting an assistant but not credential scoping controls. Missing for 10: explicit scope/permission model for API keys or OAuth tokens issued to agents, documentation of least-privilege credential issuance, and any audit/control mechanism tied to agent-specific access.

   • [claimed-docs] “Use the Canva Connect APIs to integrate Canva capabilities into your platform or your users' workflow.”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “If you're a user wanting to connect your personal AI assistant to Canva's AI Connector, see our Help Center”
   • [claimed-docs] “Private integrations can only be used by your team on a Canva Enterprise plan.”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”

   proof ↗⚑ flag

   Spline—none0/10

   Evidence shows Spline offers an MCP server for connecting AI coding tools to the editor, but there is no mention of scoped or least-privilege API credentials, API keys with permission scopes, or any access-control mechanism for issuing restricted agent credentials. Missing for 10: any documentation of credential/token scoping, permission levels, or least-privilege API key issuance for agents.

   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [claimed-docs] “The MCP server is bundled in the desktop app, so connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here —…”
   • [claimed-docs] “The **Spline MCP Server** connects AI coding tools like **Claude**, **Cursor**, **Codex**, and **Antigravity** directly to your Spline edito…”
   • [probe] “official MCP server documented at https://docs.spline.design/basics/download-spline-for-desktop”

   proof ↗⚑ flag

7. ai-native userBuild against official SDKs

   weight 2 · round to Canva
   Canva✓fullprobed8/10

   Canva provides official SDKs (Apps SDK, Connect APIs) plus an OpenAPI description enabling generated client SDKs in various languages, a CLI for app development, and an official MCP server for AI-native integration—covering the core of building against official SDKs. Missing for 10: independent/hands-on developer corroboration beyond first-party docs, and the openapi.json probe returned 404 suggesting the spec isn't trivially discoverable at expected paths.

   • [claimed-docs] “Use the Canva Apps SDK to bring new design and editing experiences into the Canva editor.”
   • [claimed-docs] “Use the Canva Connect APIs to integrate Canva capabilities into your platform or your users' workflow.”
   • [claimed-docs] “You can use this description to generate a client SDK in your preferred language, using a code generation library”
   • [claimed-docs] “Leverage a command line tool designed for creating and managing Canva apps.”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries...to generate client SDKs in vari…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://www.canva.dev/openapi.json, https://www.canva.dev/swagger.json, https://www.canva.dev/api/op…”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”
   • [probe] “official CLI documented at https://www.canva.dev/docs/apps/quickstart/”

   proof ↗⚑ flag

   Spline~partialprobed6/10

   Spline documents a Code API and multiple official code-export targets (Vanilla JS, Three.js, React, Next.js, react-three-fiber) plus a self-hosted runtime export, which functions as an SDK-like integration surface for developers (spline-docs-6, spline-docs-12, spline-docs-17, spline-docs-8). However, a live probe for a formal OpenAPI/SDK spec returned 404s, and there's no independent developer corroboration of SDK usage or versioned package docs beyond the docs site. Missing for 10: dedicated SDK/package documentation (e.g. npm SDK docs), OpenAPI/API reference, and third-party hands-on confirmation of building against these SDKs.

   • [claimed-docs] “It enables you to trigger actions, modify object properties, and create dynamic, responsive 3D environments directly from your existing code…”
   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS, Three.js, React, Next.js, react-three-fiber”
   • [claimed-docs] “The Code API allows you to seamlessly integrate interactive 3D experiences into your projects, offering greater control and customization.”
   • [claimed-docs] “The Self-Hosted export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a fully se…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.spline.design/openapi.json, https://docs.spline.design/swagger.json, https://docs.splin…”

   proof ↗⚑ flag

8. ai-native userSubscribe to events via webhooks

   weight 2 · round to Spline
   Canva—none0/10

   No evidence of any webhook/event-subscription capability in the Canva Connect APIs, MCP server, or other docs — only REST-style APIs, SCIM, audit logs, and MCP tool interactions are described. missing for 10: any mention of webhooks, event subscriptions, or push notifications for state changes.

   ⚑ flag

   Spline✓fullclaimed7/10

   Spline docs explicitly document a Webhooks feature that allows listening for and responding to external events, triggering actions, updating variables, and reacting to data sent from external sources — directly matching the story. However, this is first-party documentation only with no independent/hands-on corroboration or detail on subscription mechanics (event types, payload schema, delivery guarantees). Missing for 10: independent verification of webhook reliability, detailed event-type/payload documentation, hands-on community testing.

   • [claimed-docs] “Webhooks in Spline allow you to create real-time, event-driven interactions by listening for and responding to external events.”
   • [claimed-docs] “By using webhooks, your Spline project can automatically trigger actions, update variables, and react to data sent from external sources”
   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”

   proof ↗⚑ flag

Agentic features

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

   weight 2 · round drawn
   Canva~partialclaimed4/10

   Canva provides integration surfaces (Connect APIs, Apps SDK, Bulk create, Brand Template autofill, MCP server) that let developers or AI assistants trigger Canva actions programmatically, but these are all invoked by an external caller rather than being a built-in autonomous scheduler or background automation engine within Canva itself. Missing for 10: evidence of native scheduled/triggered automations, event-driven workflows, or an 'autopilot' feature that runs without external initiation.

   • [claimed-docs] “Use the Canva Connect APIs to integrate Canva capabilities into your platform or your users' workflow.”
   • [claimed-docs] “Automatically personalize and populate Brand Templates with real-time content from your data sources.”
   • [claimed-docs] “Scale design output - generate personalized content in bulk using your own data and visuals.”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [claimed-docs] “Import data directly from third-party sources to use for design generation or Bulk create.”

   proof ↗⚑ flag

   Spline~partialclaimed4/10

   Spline offers webhooks and a Real-time API that let a published scene 'automatically trigger actions, update variables, and react to data sent from external sources' without a person driving it (spline-docs-9, spline-docs-10, spline-docs-26), which is a form of event-driven automation running unattended. However, this is scene-level reactivity to external events, not a general-purpose scheduled/autonomous agent workflow — the AI Agent and MCP integrations require an open editor tab and active AI-tool session (spline-docs-13, spline-docs-25) rather than running independently in the background. Missing for 10: evidence of scheduled/cron-like autonomous tasks, independent confirmation of webhook automation working in practice, and any agentic workflow that persists without an open editor session.

   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “Webhooks in Spline allow you to create real-time, event-driven interactions by listening for and responding to external events.”
   • [claimed-docs] “By using webhooks, your Spline project can automatically trigger actions, update variables, and react to data sent from external sources”
   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [claimed-docs] “The MCP server is bundled in the desktop app, so connecting Claude, Cursor, VS Code, Antigravity or ChatGPT to your editor only works here —…”

   proof ↗⚑ flag

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

   weight 3 · round to Spline
   Canva~partialclaimed3/10

   Evidence mentions a 'Canva AI' feature only in passing (pasting links to add context to designs) but provides no detail on delegating broader tasks to a built-in assistant; most AI-related evidence instead describes the MCP server that lets *external* AI assistants control Canva, not an assistant embedded within Canva's own UI. missing for 10: documentation of an in-app AI assistant's task scope, examples of users delegating multi-step design tasks to it, and independent/hands-on confirmation of its capabilities.

   • [claimed-docs] “Bring content into Canva AI by pasting links from other platforms to design with more context.”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”

   proof ↗⚑ flag

   Spline✓fullclaimed7/10

   Spline ships a built-in AI Agent that autonomously builds and edits 3D scenes ('describe what you want and it does the work through the editor itself — creating objects, setting up materials and lighting, animating, wiring interactivity') and a parallel Hana AI Agent for 2D screens, both directly matching the delegated-task story. Missing for 10: independent/hands-on verification of the agent's real-world reliability and scope beyond first-party docs.

   • [claimed-docs] “The AI Agent builds and edits 3D scenes for you. Describe what you want and it does the work through the editor itself — creating objects, s…”
   • [claimed-docs] “The **AI Agent** builds and edits 3D scenes for you. Describe what you want and it does the work through the editor itself — creating object…”
   • [claimed-docs] “The AI Agent builds and edits 3D scenes for you. Describe what you want and it does the work through the editor itself”
   • [claimed-docs] “Hana's **AI Agent** designs for you. Describe a screen and it builds it on the canvas: layout, type, colour, effects, states and interaction…”
   • [claimed-docs] “Hana's AI Agent designs for you. Describe a screen and it builds it on the canvas: layout, type, colour, effects, states and interactions.”
   • [claimed-docs] “describe a screen and Hana designs it, then edit it by hand”

   proof ↗⚑ flag

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

   weight 2 · round drawn
   Canva✓fullprobed8/10

   Canva ships an official MCP server explicitly designed so AI assistants can operate Canva via natural language — supporting design generation, editing, discovery, asset/brand management, export, and comments through NL commands, and this is corroborated by a probe confirming the documented endpoint. This directly satisfies the AI-native natural-language operation story, though it relies on third-party AI clients rather than a native in-product NL interface. Missing for 10: independent/hands-on user reports of NL command reliability, and evidence of a first-party conversational interface within Canva itself.

   • [claimed-docs] “give your users access to Canva's design capabilities through natural language, directly from the AI tools they already use”
   • [claimed-docs] “Design generation: Create designs programmatically from text descriptions and specifications”
   • [claimed-docs] “Design editing: Apply targeted modifications to existing designs through natural language commands”
   • [claimed-docs] “Design discovery: Search and retrieve designs, pages, and folders from your design library”
   • [claimed-docs] “Asset and brand management: Upload and organize assets, manage folders, and access brand kits and templates”
   • [claimed-docs] “Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “Collaborative workflows: Add comments to designs and integrate feedback loops directly into your workflows”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”

   proof ↗⚑ flag

   Spline✓fullclaimed8/10

   Spline's AI Agent lets users describe what they want in natural language and the editor builds objects, materials, lighting, animation, and interactivity accordingly (spline-docs-3, spline-docs-20, spline-docs-40), and Hana's AI Agent similarly designs UI screens from natural-language descriptions (spline-docs-21, spline-docs-39, spline-docs-54). AI texture generation also works by typing descriptions (spline-docs-4, spline-docs-28). This is reinforced by an MCP server enabling natural-language driven control from external AI clients like Claude/Cursor (spline-docs-13, spline-docs-36). Missing for 10: independent/hands-on user reports confirming natural-language command reliability and breadth of coverage beyond first-party docs.

   • [claimed-docs] “describe what you want and the 3D editor builds it: objects, materials, lighting, animation, interactivity.”
   • [claimed-docs] “The AI Agent builds and edits 3D scenes for you. Describe what you want and it does the work through the editor itself — creating objects, s…”
   • [claimed-docs] “describe a screen and Hana designs it, then edit it by hand”
   • [claimed-docs] “Hana's **AI Agent** designs for you. Describe a screen and it builds it on the canvas: layout, type, colour, effects, states and interaction…”
   • [claimed-docs] “The **AI Agent** builds and edits 3D scenes for you. Describe what you want and it does the work through the editor itself — creating object…”
   • [claimed-docs] “Use Artificial Intelligence to generate textures like grass, concrete, wood, ceramic, and patterns just by typing!”
   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [claimed-docs] “The **Spline MCP Server** connects AI coding tools like **Claude**, **Cursor**, **Codex**, and **Antigravity** directly to your Spline edito…”

   proof ↗⚑ flag

Api quality

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

   weight 2 · round drawn
   Canva—none0/10

   Canva documents an OpenAPI description for its Connect APIs (used for generating client SDKs) but there is no evidence of an interactive API reference with runnable/try-it examples; probes for openapi.json/swagger endpoints returned 404s, suggesting no live interactive console is exposed.

   • [claimed-docs] “You can use this description to generate a client SDK in your preferred language, using a code generation library”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries...to generate client SDKs in vari…”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries... to generate client SDKs in var…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://www.canva.dev/openapi.json, https://www.canva.dev/swagger.json, https://www.canva.dev/api/op…”

   proof ↗⚑ flag

   Spline—none0/10

   Spline documents several APIs (Code API, Real-time API, webhooks) but the evidence pack shows no interactive API reference with runnable examples — a probe for OpenAPI/swagger specs returned 404 on all candidate paths, and the docs are plain markdown descriptions rather than an interactive sandbox.

   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.spline.design/openapi.json, https://docs.spline.design/swagger.json, https://docs.splin…”
   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “The Code API allows you to seamlessly integrate interactive 3D experiences into your projects, offering greater control and customization.”

   proof ↗⚑ flag

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

   weight 2 · round to Canva
   Canva✓fullprobed7/10

   Canva's developer docs explicitly state it provides an OpenAPI description of the Connect APIs that can be used with code-generation libraries to produce client SDKs, directly matching the story. Confidence is tempered because the probe couldn't locate the actual spec file at common guessed paths (only generic /openapi.json etc. were tried, not the real documented location), so independent verification of downloadability is thin. Missing for 10: a probed/verified working download URL for the OpenAPI file and independent (non-vendor) confirmation it's actually fetchable.

   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries...to generate client SDKs in vari…”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries... to generate client SDKs in var…”
   • [claimed-docs] “You can use this description to generate a client SDK in your preferred language, using a code generation library”
   • [probe] “PROBE openapi: all candidate paths 404 (https://www.canva.dev/openapi.json, https://www.canva.dev/swagger.json, https://www.canva.dev/api/op…”

   proof ↗⚑ flag

   Spline—none0/10

   A direct probe for OpenAPI/swagger specs at common paths returned 404 across all candidates, and no documentation references a downloadable machine-readable API spec; Spline offers a Real-time API and Code API but no OpenAPI equivalent is evidenced.

   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.spline.design/openapi.json, https://docs.spline.design/swagger.json, https://docs.splin…”

   proof ↗⚑ flag

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

   weight 2 · round drawn
   Canva—none0/10

   Evidence shows Canva has Connect APIs, an OpenAPI description, SDKs, and an MCP server, but there is no mention anywhere of API versioning scheme or a documented deprecation policy for these APIs; the OpenAPI spec itself isn't even discoverable at expected endpoints (canva-probe-3). missing for 10: explicit API versioning scheme, documented deprecation/sunset policy, changelog or migration guides for breaking changes.

   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries...to generate client SDKs in vari…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://www.canva.dev/openapi.json, https://www.canva.dev/swagger.json, https://www.canva.dev/api/op…”

   proof ↗⚑ flag

   Spline—none0/10

   No evidence of API versioning scheme or a documented deprecation policy; the OpenAPI probe returned 404s across all candidate paths, and no docs mention version numbers, changelogs, or deprecation timelines for the Real-time API, Code API, or MCP server.

   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.spline.design/openapi.json, https://docs.spline.design/swagger.json, https://docs.splin…”

   proof ↗⚑ flag

Ai

1. designerBuilt-in AI handles asset chores — image generation, background removal, content-aware fills, copy suggestions — inside the editor

   weight 1 · round to Spline
   Canva—none0/10

   The evidence pack only covers developer-facing surfaces (Apps SDK, Connect APIs, MCP server, Admin/SCIM/audit APIs) and never mentions Canva's native in-editor AI asset tools like Magic Media image generation, Magic Eraser/background removal, Magic Edit content-aware fill, or Magic Write copy suggestions used directly by a designer inside the editor. Missing for 10: any documentation of built-in Magic Studio-style tools (image generation, background removal, content-aware fill, copy suggestions) operating natively within the Canva editor UI.

   ⚑ flag

   Spline~partialclaimed3/10

   Spline documents built-in AI features—AI-generated textures via text prompts and an AI Agent/Hana that builds/edits 3D scenes and screens—which is a form of in-editor generative asset creation, but there's no evidence of background removal, content-aware fill, or AI copy/text suggestions inside the editor. missing for 10: background removal, content-aware fill, copy/text suggestions, general image asset generation beyond textures.

   • [claimed-docs] “Use Artificial Intelligence to generate textures like grass, concrete, wood, ceramic, and patterns just by typing!”
   • [claimed-docs] “Use Artificial Intelligence to generate textures like **grass**, **concrete**, **wood**, **ceramic**, and **patterns** just by typing!”
   • [claimed-docs] “The AI Agent builds and edits 3D scenes for you. Describe what you want and it does the work through the editor itself — creating objects, s…”
   • [claimed-docs] “describe a screen and Hana designs it, then edit it by hand”

   proof ↗⚑ flag

2. designerGenerate or edit designs from a natural-language prompt — first drafts, layouts, or full mockups the tool actually produces

   weight 2 · round to Spline
   Canva~partialprobed6/10

   Canva's docs describe design generation and editing from natural language via its MCP server (canva-docs-17, canva-docs-18, canva-docs-28, canva-docs-33), which lets AI assistants create and modify designs from text descriptions and specifications, plus discovery, export, and asset management. However, this capability is delivered through an external MCP integration/AI connector rather than a native in-editor 'Magic Design'-style prompt feature, and there's no independent/hands-on evidence confirming quality of generated mockups. Missing for 10: first-party in-product documentation of native prompt-to-design (e.g., Magic Design) generation, independent hands-on verification of output quality, and confirmation this works without third-party AI tool setup.

   • [claimed-docs] “Design generation: Create designs programmatically from text descriptions and specifications”
   • [claimed-docs] “Design editing: Apply targeted modifications to existing designs through natural language commands”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”
   • [claimed-docs] “Design generation:** Create designs programmatically from text descriptions and specifications”
   • [claimed-docs] “Design discovery:** Search and retrieve designs, pages, and folders from your design library”
   • [claimed-docs] “Design export:** Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”

   proof ↗⚑ flag

   Spline✓fullclaimed7/10

   Spline documents two AI Agents that generate designs from natural-language prompts: the 3D AI Agent creates objects, materials, lighting, animation and interactivity from a description, and Hana's AI Agent designs full 2D screens (layout, type, colour, effects, states) from a text description, which can then be edited by hand. This directly matches the story of generating first drafts/mockups from a prompt. Missing for 10: independent/hands-on verification of the AI Agent's actual output quality, and no third-party reviews or demos confirming real-world reliability of these AI generation features.

   • [claimed-docs] “The AI Agent builds and edits 3D scenes for you. Describe what you want and it does the work through the editor itself — creating objects, s…”
   • [claimed-docs] “describe a screen and Hana designs it, then edit it by hand”
   • [claimed-docs] “Hana's **AI Agent** designs for you. Describe a screen and it builds it on the canvas: layout, type, colour, effects, states and interaction…”
   • [claimed-docs] “The **AI Agent** builds and edits 3D scenes for you. Describe what you want and it does the work through the editor itself — creating object…”
   • [claimed-docs] “Hana's AI Agent designs for you. Describe a screen and it builds it on the canvas: layout, type, colour, effects, states and interactions.”
   • [claimed-docs] “The AI Agent builds and edits 3D scenes for you. Describe what you want and it does the work through the editor itself”

   proof ↗⚑ flag

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

   weight 2 · round to Canva
   Canva✓fullclaimed7/10

   Canva explicitly supports bulk operations via Bulk Create with data import, personalized Brand Template population at scale, and Connect APIs to sync/create many designs/assets programmatically, plus MCP/API access for AI-native automation across many items. Missing for 10: independent/hands-on verification of scale limits, no explicit rate-limit or batch-size documentation, and no third-party confirmation of true 'many items' throughput.

   • [claimed-docs] “Automatically personalize and populate Brand Templates with real-time content from your data sources.”
   • [claimed-docs] “Scale design output - generate personalized content in bulk using your own data and visuals.”
   • [claimed-docs] “Import data directly from third-party sources to use for design generation or Bulk create.”
   • [claimed-docs] “You can use the Connect APIs to create web app integrations that create and sync assets, designs, comments, and more.”
   • [claimed-docs] “The Canva MCP server enables AI assistants to seamlessly interact with Canva's design capabilities.”

   proof ↗⚑ flag

   Spline—none0/10

   Spline's docs describe single-scene editing, MCP/Code API/webhook integrations, and AI generation features, but nothing indicates batch or bulk processing across many objects, scenes, or files at once (e.g., no scripting loop API, no bulk export/import, no multi-file automation mentioned).

   ⚑ flag

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

   weight 3 · round to Spline
   Canva—none0/10

   Canva's evidence covers APIs, apps, MCP server, bulk creation, and integrations, but nothing describes user-defined rule/trigger automation (event-condition-action logic) that fires actions automatically on events. No workflow/automation-rule engine, webhooks-as-triggers, or 'when X happens do Y' capability is documented.

   ⚑ flag

   Spline✓fullclaimed7/10

   Spline's states/events/actions system, webhooks, and Real-time API let users define rules that trigger actions automatically on events (e.g., 'automatically trigger actions, update variables, and react to data sent from external sources'), which is core automation-depth functionality. missing for 10: independent/hands-on verification of event-rule automation in practice, and more detail on the breadth/complexity of conditional logic supported beyond docs claims.

   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “Webhooks in Spline allow you to create real-time, event-driven interactions by listening for and responding to external events.”
   • [claimed-docs] “By using webhooks, your Spline project can automatically trigger actions, update variables, and react to data sent from external sources”
   • [claimed-docs] “Most interactivity in Spline needs no code at all. It's built from **states, events and actions**”

   proof ↗⚑ flag

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

   weight 1 · round to Spline
   Canva—none0/10

   Evidence shows audit logs tracking user activity, but no documentation of versioning, review, or rollback specifically for automations (Connect API integrations, MCP workflows, or bulk create jobs). There's no mention of automation history, diffing, approval workflows, or reverting automated changes.

   • [claimed-docs] “Use audit logs to track changes made by your team in Canva.”
   • [claimed-docs] “Canva provides audit logs that contain records of user activities in Canva, such as installing a Canva App, exporting a design for download,…”

   proof ↗⚑ flag

   Spline~partialclaimed5/10

   Spline's Version History lets users preview, recover, and manually save previous file states, which would cover the interactivity/automation logic (events, actions, webhooks) embedded in a scene, but this is a whole-file history feature, not a dedicated review/diff mechanism for automations specifically, and no evidence of granular change review or approval workflows exists. missing for 10: automation-specific diffing/review UI, branching or approval workflow, independent confirmation of rollback reliability for complex automations.

   • [claimed-docs] “Version History allows you to access and preview all the saved versions of your files. This feature helps you track changes, recover previou…”
   • [claimed-docs] “Version History allows you to access and preview all the saved versions of your files.”
   • [claimed-docs] “Code & Self-hosted exports * File version history * Team Library”

   proof ↗⚑ flag

Collaboration

1. product managerComment directly on designs, mention teammates, and resolve threads so review happens in the file, not in email

   weight 2 · round to Canva
   Canva~partialclaimed5/10

   Evidence confirms Canva supports commenting on designs and even integrating comments into external workflows (canva-docs-9, canva-docs-22, canva-docs-37), implying in-file review is possible, but there is no direct documentation of @mentioning teammates or explicitly 'resolving' comment threads within the editor. missing for 10: explicit documentation of @mention functionality, thread resolution/closing feature, and any first-party or independent confirmation of the full in-app review workflow.

   • [claimed-docs] “Receive and reply to Canva comments from any platform and simplify team communication.”
   • [claimed-docs] “Collaborative workflows: Add comments to designs and integrate feedback loops directly into your workflows”
   • [claimed-docs] “You can use the Connect APIs to create web app integrations that create and sync assets, designs, comments, and more.”

   proof ↗⚑ flag

   Spline~partialclaimed4/10

   Docs confirm real-time multiplayer editing and that users can 'give feedback in place, the way they would in a 2D design tool,' implying some in-file commenting, but there is no mention of @mentioning teammates or resolving comment threads. missing for 10: explicit mention/tagging of teammates, ability to resolve/close comment threads, dedicated commenting UI documentation.

   • [claimed-docs] “Several people can edit the same file at the same time and give feedback in place, the way they would in a 2D design tool.”
   • [claimed-docs] “It's multiplayer. Several people can edit the same file at the same time and give feedback in place, the way they would in a 2D design tool.”
   • [claimed-docs] “is multiplayer from the ground up: your files live in the cloud and several people can edit the same one at once.”

   proof ↗⚑ flag

2. designerMy whole team can edit the same file simultaneously with live cursors and instant sync

   weight 3 · round to Spline
   Canva—none0/10

   The evidence pack focuses entirely on developer APIs, SDKs, MCP servers, and admin/provisioning tools; there is no mention of real-time multi-user editing, live cursors, or simultaneous in-editor collaboration features that are core to this story.

   ⚑ flag

   Spline✓fullclaimed8/10

   Multiple first-party docs consistently describe Spline as "multiplayer from the ground up," cloud-hosted files enabling several people to edit the same file simultaneously with in-place feedback, similar to 2D design tools — directly matching the story. Missing for 10: explicit mention of live cursors specifically, and no independent/hands-on corroboration of real-time sync quality at scale.

   • [claimed-docs] “is multiplayer from the ground up: your files live in the cloud and several people can edit the same one at once”
   • [claimed-docs] “is multiplayer from the ground up: your files live in the cloud and several people can edit the same one at once.”
   • [claimed-docs] “Several people can edit the same file at the same time and give feedback in place, the way they would in a 2D design tool.”
   • [claimed-docs] “It's multiplayer. Several people can edit the same file at the same time and give feedback in place, the way they would in a 2D design tool.”
   • [claimed-docs] “Spline is two editors that share the same interactivity model, the same real-time collaboration, and the same publishing pipeline.”

   proof ↗⚑ flag

Design systems

1. designerManage design tokens or variables — colors, spacing, typography — with modes or themes applied across designs

   weight 2 · round drawn
   Canva—none0/10

   Canva's evidence covers Brand Templates, Apps SDK, Connect APIs, and MCP integrations, but nothing describes design tokens/variables (colors, spacing, typography) as reusable, mode-based theme objects applied programmatically across designs — Canva's brand kit concept is not documented as a token/variable system with theme modes.

   ⚑ flag

   Spline—none0/10

   Spline's docs mention 'variables and data' only in the context of real-time API/webhooks event data, not design tokens for color, spacing or typography, and the only 'theme' feature found is the editor's own Dark/Light/System UI theme, not a token system applied across a designer's own scenes. No evidence shows reusable style tokens with modes/themes propagated across designs.

   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “Webhooks in Spline allow you to create real-time, event-driven interactions by listening for and responding to external events.”
   • [claimed-docs] “By using webhooks, your Spline project can automatically trigger actions, update variables, and react to data sent from external sources”
   • [claimed-docs] “Spline supports **Dark**, **Light**, and **System** themes so you can customize the interface to your preference.”

   proof ↗⚑ flag

2. designerBuild reusable components with variants and per-instance overrides that stay linked to their source

   weight 3 · round to Spline
   Canva—none0/10

   The evidence pack covers Canva's APIs, SDKs, MCP server, and Brand Templates but contains no mention of a component system with variants, per-instance overrides, or linked source instances (like Figma-style components). Brand Templates allow personalization but are not described as variant-based linked components. Missing for 10: any documentation of a component/variant system, linked instance overrides, or source-of-truth syncing mechanism for design elements.

   ⚑ flag

   Spline~partialclaimed4/10

   Spline docs confirm a Components system where reusable objects propagate changes from a main component to all instances (spline-docs-5), which covers the 'linked to source' part of the story. However, there is no evidence of a variants system (multiple named states of a component) or per-instance override capability that persists while staying linked, as seen in dedicated design tools. Missing for 10: documented variants/property sets per component, explicit per-instance override UI, and evidence of overrides surviving source updates.

   • [claimed-docs] “Components are objects that can be reused multiple times throughout your scene and whenever there are changes made to the main component all…”

   proof ↗⚑ flag

3. designerShared libraries publish components and styles across files and projects, with controlled updates when the source changes

   weight 2 · round drawn
   Canva~partialclaimed4/10

   Canva offers Brand Kits and Brand Templates that centralize on-brand assets/styles for teams to reuse across designs (canva-docs-15, canva-docs-7, canva-docs-20), which is the closest analog to a shared component library. However, there is no evidence of a true library-linking mechanism where edits to a source style/component automatically or optionally propagate updates to files that already used it, which is the core of the story. missing for 10: documentation of update propagation/versioning when a Brand Template or Brand Kit element changes, evidence of per-instance 'update available' notifications, independent confirmation of this workflow in practice.

   • [claimed-docs] “Make Brand Templates easily accessible, so teams can quickly create and edit on-brand designs.”
   • [claimed-docs] “Automatically personalize and populate Brand Templates with real-time content from your data sources.”
   • [claimed-docs] “Asset and brand management: Upload and organize assets, manage folders, and access brand kits and templates”

   proof ↗⚑ flag

   Spline~partialclaimed4/10

   Spline documents component reuse where instances update when the main component changes (spline-docs-5), and pricing mentions a 'Team Library' feature (spline-docs-33), implying some shared-library capability across a team. However, there's no documentation detailing publishing components/styles across separate files or projects, or how updates propagate/are controlled at that scope. Missing for 10: cross-file/cross-project library publishing workflow, versioned/controlled update mechanism, and any independent confirmation of Team Library behavior.

   • [claimed-docs] “Components are objects that can be reused multiple times throughout your scene and whenever there are changes made to the main component all…”
   • [claimed-docs] “Code & Self-hosted exports * File version history * Team Library”

   proof ↗⚑ flag

Handoff

1. developerDesigns map to production code — component-to-code mappings or direct design-to-code output I can actually ship

   weight 2 · round to Spline
   Canva—none0/10

   Canva's developer docs cover Apps SDK, Connect APIs, MCP server, and CLI for building integrations, but none of the evidence describes design-to-code output or component-to-code mappings that map a Canva design to shippable UI code (e.g., React/CSS export) as design tools like Figma do.

   • [claimed-docs] “Use the Canva Apps SDK to bring new design and editing experiences into the Canva editor.”
   • [claimed-docs] “Use the Canva Connect APIs to integrate Canva capabilities into your platform or your users' workflow.”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [probe] “official MCP server documented at https://www.canva.dev/docs/mcp/”
   • [probe] “official CLI documented at https://www.canva.dev/docs/apps/quickstart/”

   proof ↗⚑ flag

   Spline✓fullclaimed7/10

   Spline documents direct design-to-code export in multiple production-ready formats (Vanilla JS, Three.js, React, Next.js, react-three-fiber), plus a Code API for wiring exported scenes into existing codebases, a self-hosted ZIP export bundling assets/runtime, and a native <spline-viewer> web component for drop-in embedding — all clearly aimed at shipping designed 3D/2D scenes into real production code. Missing for 10: independent/hands-on verification that exported code is production-quality and maintainable, and explicit component-to-code sync (e.g., design system tokens/components mapped 1:1 to code components) beyond scene-level export.

   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS, Three.js, React, Next.js, react-three-fiber”
   • [claimed-docs] “Spline allows you to export the code of your scene as: **Vanilla JS**... **Three.js**... **React**... **Next.js**... **react-three-fiber**”
   • [claimed-docs] “Spline allows you to export the code of your scene as: * **Vanilla JS**”
   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS ... Three.js ... React ... Next.js ... react-three-fiber”
   • [claimed-docs] “It enables you to trigger actions, modify object properties, and create dynamic, responsive 3D environments directly from your existing code…”
   • [claimed-docs] “The Code API allows you to seamlessly integrate interactive 3D experiences into your projects, offering greater control and customization.”
   • [claimed-docs] “It enables you to trigger actions, modify object properties, and create dynamic, responsive 3D environments directly from your existing code”
   • [claimed-docs] “The Self-Hosted export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a fully se…”
   • [claimed-docs] “allows you to export your entire project, including all assets and the Spline runtime, as a fully self-contained ZIP file”
   • [claimed-docs] “The **Self-Hosted** export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a full…”
   • [claimed-docs] “`<spline-viewer>` is a native HTML web component and can be used with most web builders, frameworks, or browsers.”

   proof ↗⚑ flag

2. developerInspect any layer and get measurements, styles, assets, and code snippets (CSS/iOS/Android) without an editor seat

   weight 3 · round drawn
   Canva—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

   Spline—none0/10

   Spline's docs describe code export as full-scene exports (Vanilla JS, Three.js, React, Next.js, react-three-fiber) and a Code API for developers, but there is no evidence of a dev-handoff/inspect mode that lets a non-editor viewer inspect a specific layer for measurements, styles, assets, or CSS/iOS/Android snippets without holding an editor seat.

   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS, Three.js, React, Next.js, react-three-fiber”
   • [claimed-docs] “The Code API allows you to seamlessly integrate interactive 3D experiences into your projects, offering greater control and customization.”
   • [claimed-docs] “Spline allows you to export the code of your scene as: **Vanilla JS**... **Three.js**... **React**... **Next.js**... **react-three-fiber**”
   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS ... Three.js ... React ... Next.js ... react-three-fiber”

   proof ↗⚑ flag

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

   weight 2 · round drawn
   Canva~partialclaimed5/10

   Canva offers substantial API surface (Connect APIs, Apps SDK, SCIM, Admin API, MCP server) covering design creation, editing, export, comments, and asset management, but this is a curated set of endpoints rather than full parity with the UI editor's rich manual design capabilities (e.g. free-form drag-and-drop editing, most granular design tools). No evidence confirms full UI-API parity; the docs describe specific supported operations, not a claim of complete equivalence. Missing for 10: explicit parity claim or evidence that every UI editing action (e.g. fine-grained manual layout, all effects/filters) is API-accessible, and independent verification of coverage limits.

   • [claimed-docs] “Use the Canva Connect APIs to integrate Canva capabilities into your platform or your users' workflow.”
   • [claimed-docs] “Design generation: Create designs programmatically from text descriptions and specifications”
   • [claimed-docs] “Design editing: Apply targeted modifications to existing designs through natural language commands”
   • [claimed-docs] “Design discovery: Search and retrieve designs, pages, and folders from your design library”
   • [claimed-docs] “Asset and brand management: Upload and organize assets, manage folders, and access brand kits and templates”
   • [claimed-docs] “Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “Collaborative workflows: Add comments to designs and integrate feedback loops directly into your workflows”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “You can use the Connect APIs to create web app integrations that create and sync assets, designs, comments, and more.”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries...to generate client SDKs in vari…”

   proof ↗⚑ flag

   Spline~partialprobed5/10

   Spline offers real API-like surfaces for AI-native workflows — a Code API for runtime object/property control (spline-docs-6, -17), a Real-time API and webhooks for external data/events (spline-docs-9, -10, -26), and notably an official MCP server that lets AI clients like Claude/Cursor/Codex drive the live editor tab directly (spline-docs-13, -36, -53) plus an AI Agent that can create objects, materials, lighting and interactivity through the editor (spline-docs-20, -40, -55). However, probing for a formal REST/OpenAPI spec returned 404s across all candidate paths (spline-probe-3), and the documented APIs are scoped to runtime manipulation/events, not full parity with the UI's modeling toolset (booleans, hair systems, physics, gaussian splatting, materials library) — so full UI-equivalent API coverage is not evidenced. missing for 10: a documented comprehensive API/SDK exposing full scene-authoring parity (modeling, materials, physics, animation) rather than just runtime triggers/MCP-driven editor control, and independent confirmation of API completeness.

   • [claimed-docs] “It enables you to trigger actions, modify object properties, and create dynamic, responsive 3D environments directly from your existing code…”
   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “Webhooks in Spline allow you to create real-time, event-driven interactions by listening for and responding to external events.”
   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [claimed-docs] “The AI Agent builds and edits 3D scenes for you. Describe what you want and it does the work through the editor itself — creating objects, s…”
   • [claimed-docs] “The **Spline MCP Server** connects AI coding tools like **Claude**, **Cursor**, **Codex**, and **Antigravity** directly to your Spline edito…”
   • [claimed-docs] “The Spline MCP Server connects AI coding tools like Claude, Cursor, Codex, and Antigravity directly to your Spline editor.”
   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.spline.design/openapi.json, https://docs.spline.design/swagger.json, https://docs.splin…”

   proof ↗⚑ flag

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

   weight 3 · round to Spline
   Canva~partialclaimed4/10

   Canva supports exporting individual designs in open, standard formats (PDF, PNG, JPG, PPTX, MP4) via both the Connect API and the MCP server, giving users a way to pull their design content out in interoperable formats. However there is no evidence of a comprehensive account-level data export (brand kits, comments, team assets, settings, history) or a documented 'export all my data' / account deletion-and-portability feature that would let an AI-native user fully leave with everything intact. missing for 10: bulk/account-wide data export, documented data portability or GDPR-style export tool, evidence of exporting non-design data (comments, brand assets, admin/audit data) in open formats.

   • [claimed-docs] “Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “Design export:** Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “You can use the Connect APIs to create web app integrations that create and sync assets, designs, comments, and more.”

   proof ↗⚑ flag

   Spline✓fullclaimed7/10

   Spline documents self-hosted export (full project, assets, and runtime as a self-contained ZIP) and code export in multiple open web formats (Vanilla JS, Three.js, React, Next.js, react-three-fiber), giving users a way to take their scene data out of the platform entirely rather than being locked into a proprietary cloud-only format. Missing for 10: no evidence of exporting to standard interchange 3D formats (e.g., glTF/OBJ), no independent/hands-on confirmation that self-hosted exports are truly complete and portable, and no mention of exporting non-visual metadata (variables, version history) alongside the scene.

   • [claimed-docs] “The Self-Hosted export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a fully se…”
   • [claimed-docs] “allows you to export your entire project, including all assets and the Spline runtime, as a fully self-contained ZIP file”
   • [claimed-docs] “The **Self-Hosted** export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a full…”
   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS, Three.js, React, Next.js, react-three-fiber”
   • [claimed-docs] “Spline allows you to export the code of your scene as: **Vanilla JS**... **Three.js**... **React**... **Next.js**... **react-three-fiber**”
   • [claimed-docs] “Code & Self-hosted exports * File version history * Team Library”

   proof ↗⚑ flag

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

   weight 2 · round drawn
   Canva—none0/10

   Canva is closed-source SaaS; the evidence pack only covers APIs, SDKs, MCP server, and developer docs — none of it indicates any open-license availability of Canva's own product source code.

   ⚑ flag

   Spline—none0/10

   Spline is a closed, proprietary commercial design tool; nothing in the evidence pack mentions an open-source license or publicly available source code for the core product (only exported project assets can be self-hosted, which is not the same as open-sourcing the tool's source).

   ⚑ flag

4. ai-native userSelf-host the core product

   weight 3 · round drawn
   Canva—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

   Spline—none0/10

   Spline is an inherently cloud/multiplayer SaaS editor ('files live in the cloud', 'multiplayer from the ground up') with no evidence of a self-hostable core editor. The 'Self-Hosted' feature only lets users export a finished scene/runtime as a static ZIP for embedding elsewhere — it does not let anyone run the Spline design tool itself on their own infrastructure.

   • [claimed-docs] “is multiplayer from the ground up: your files live in the cloud and several people can edit the same one at once”
   • [claimed-docs] “is multiplayer from the ground up: your files live in the cloud and several people can edit the same one at once.”
   • [claimed-docs] “The Self-Hosted export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a fully se…”
   • [claimed-docs] “allows you to export your entire project, including all assets and the Spline runtime, as a fully self-contained ZIP file”
   • [claimed-docs] “The **Self-Hosted** export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a full…”
   • [claimed-docs] “Spline runs in the browser on any platform, with nothing to install and no GPU requirements to check.”

   proof ↗⚑ flag

Plugins

1. developerA documented plugin API with typings and samples lets me build and distribute my own editor extensions

   weight 2 · round to Canva
   Canva✓fullprobed8/10

   Canva provides a documented Apps SDK for building editor extensions, a CLI for creating/managing apps, and distribution via public Apps Marketplace or private team sharing, matching the developer story closely. missing for 10: independent/hands-on corroboration of typings quality and sample completeness, and a working OpenAPI/typings endpoint (probe found openapi.json paths returning 404).

   • [claimed-docs] “Use the Canva Apps SDK to bring new design and editing experiences into the Canva editor.”
   • [claimed-docs] “Leverage a command line tool designed for creating and managing Canva apps.”
   • [claimed-docs] “You can use the Apps SDK to create apps that import content into Canva, add elements to a user's design, automate common design tasks, and m…”
   • [claimed-docs] “You can then either distribute the app publicly via the Apps Marketplace, or share it privately with members of your team.”
   • [claimed-docs] “Team apps are created solely for a team using Canva, such as an app created to integrate into a specific system.”
   • [probe] “official CLI documented at https://www.canva.dev/docs/apps/quickstart/”
   • [probe] “PROBE openapi: all candidate paths 404 (https://www.canva.dev/openapi.json, https://www.canva.dev/swagger.json, https://www.canva.dev/api/op…”

   proof ↗⚑ flag

   Spline—none0/10

   Spline documents a Code API, MCP server, webhooks, and export pipelines for integrating exported scenes into external apps, but there is no evidence of a documented plugin/extension API (with typings and samples) for building and distributing extensions to the Spline editor itself, and an OpenAPI/SDK probe returned 404s. Missing for 10: a documented editor plugin SDK, typings/type definitions for plugin development, and sample plugin code or a plugin marketplace/distribution mechanism.

   • [claimed-docs] “It enables you to trigger actions, modify object properties, and create dynamic, responsive 3D environments directly from your existing code…”
   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “Webhooks in Spline allow you to create real-time, event-driven interactions by listening for and responding to external events.”
   • [claimed-docs] “Spline MCP Server — drive Spline from Claude, Cursor, Codex and other MCP clients, straight into your open editor tab.”
   • [probe] “PROBE openapi: all candidate paths 404 (https://docs.spline.design/openapi.json, https://docs.spline.design/swagger.json, https://docs.splin…”

   proof ↗⚑ flag

2. designerExtend the editor from a large plugin community or marketplace — icons, data fillers, accessibility checkers, and more

   weight 2 · round to Canva
   Canva~partialprobed6/10

   Canva has a real extensibility ecosystem: the Apps SDK/CLI for building apps, an Apps Marketplace for public distribution, and team-private app sharing — directly supporting icon/data/accessibility-style plugins built by third parties. However, evidence is entirely first-party docs describing the SDK/marketplace mechanics; there's no independent evidence of marketplace size, breadth of existing third-party plugins (icons, accessibility checkers, data fillers), or community adoption. Missing for 10: independent verification of marketplace scale/quality, concrete examples of third-party apps in categories named in the story, user reviews or hands-on discovery of the marketplace.

   • [claimed-docs] “You can use the Apps SDK to create apps that import content into Canva, add elements to a user's design, automate common design tasks, and m…”
   • [claimed-docs] “You can then either distribute the app publicly via the Apps Marketplace, or share it privately with members of your team.”
   • [claimed-docs] “Make apps for your team and choose who can use them.”
   • [claimed-docs] “Leverage a command line tool designed for creating and managing Canva apps.”
   • [claimed-docs] “Team apps are created solely for a team using Canva, such as an app created to integrate into a specific system.”
   • [claimed-docs] “Learn how you can turn your app into a business.”
   • [probe] “official CLI documented at https://www.canva.dev/docs/apps/quickstart/”

   proof ↗⚑ flag

   Spline—none0/10

   No evidence of a plugin community, marketplace, or third-party extensions (icons, data fillers, accessibility checkers) for Spline; the evidence only covers Spline's own built-in features (AI tools, MCP server, code export, materials library).

   ⚑ flag

Pricing

1. product managerPublished plan pricing makes seat types, tiers, and AI-feature gating clear before procurement

   weight 1 · round to Spline
   Canva—none0/10

   The evidence pack is entirely developer/API documentation (Connect APIs, Apps SDK, MCP, SCIM, Admin API) and contains no pricing page, plan comparison, seat-type breakdown, or AI-feature gating disclosure that a procurement-focused product manager would need. Only one mention ('Private integrations can only be used by your team on a Canva Enterprise plan') hints at plan gating but is far from a published pricing/tier matrix.

   • [claimed-docs] “Private integrations can only be used by your team on a Canva Enterprise plan.”

   proof ↗⚑ flag

   Spline~partialclaimed2/10

   Only one evidence item references the pricing page at all, and it's a bare fragment listing a few feature bullets (code/self-hosted export, version history, team library) with no visible seat types, tier names, or explicit AI-feature gating breakdown. Missing for 10: documented tier names/pricing, seat-type definitions, explicit AI feature gating per tier, and any independent corroboration of the pricing page's clarity.

   • [claimed-docs] “Code & Self-hosted exports * File version history * Team Library”

   proof ↗⚑ flag

2. designerThe free tier supports real design work — usable editor and collaboration without a paid seat

   weight 2 · round to Spline
   Canva—none0/10

   The evidence pack is entirely developer/API-focused (Connect APIs, Apps SDK, MCP server, SCIM, audit logs, admin API) and contains nothing about the consumer free-tier editor, its feature set, or free collaboration capabilities; one entry even implies enterprise-only integrations. No evidence supports or refutes what the free tier includes for a designer doing real design work.

   ⚑ flag

   Spline~partialclaimed3/10

   Docs describe Spline as 'multiplayer from the ground up' with cloud-based files that several people can edit at once, suggesting collaboration is core to the product rather than gated behind a paid seat, but the evidence pack has no explicit free-tier pricing page detail confirming seat limits, viewer/editor caps, or whether real-time collaboration is unrestricted on the free plan. The one pricing-page snippet (spline-docs-33) lists Code/self-hosted exports, version history, and Team Library without clarifying tier boundaries. missing for 10: explicit free-tier plan page detailing seat/collaborator limits, independent user reports confirming free-tier collaboration works in practice.

   • [claimed-docs] “is multiplayer from the ground up: your files live in the cloud and several people can edit the same one at once”
   • [claimed-docs] “is multiplayer from the ground up: your files live in the cloud and several people can edit the same one at once.”
   • [claimed-docs] “Code & Self-hosted exports * File version history * Team Library”

   proof ↗⚑ flag

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

   weight 2 · round drawn
   Canva—none0/10

   No evidence in the pack mentions data residency, regional storage options, or location controls for Canva data; documentation covers APIs, SDKs, audit logs, and provisioning but not data residency/region choice.

   ⚑ flag

   Spline—none0/10

   No evidence in the pack mentions data residency, region selection, or storage location controls; Spline docs discuss cloud storage, multiplayer, exports and MCP but never a data-region choice. missing for 10: any mention of region/residency options, data center choice, or compliance certifications tied to storage location.

   ⚑ flag

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

   weight 3 · round drawn
   Canva—none0/10

   The evidence pack contains only Canva developer docs about APIs, SDKs, MCP servers, and admin/audit tools — nothing addresses AI training data usage policies or opt-out controls for user content. Missing for 10: any documentation on AI training data policy, opt-out settings, or privacy controls related to model training.

   ⚑ flag

   Spline—none0/10

   No evidence in the pack addresses AI-training data opt-out, privacy settings, or data usage policies for model training; the docs focus entirely on design/collaboration/AI-generation features with no mention of privacy controls or training data exclusion.

   ⚑ flag

3. ai-native userControl data retention and deletion

   weight 2 · round drawn
   Canva—none0/10

   Evidence covers audit logs, SCIM provisioning, and admin/team management APIs, but nothing addresses data retention policies or user-initiated data deletion controls for AI-native usage (e.g., MCP/AI connector data handling, deletion of AI-processed content, retention settings).

   ⚑ flag

   Spline—none0/10

   The evidence pack contains no mention of data retention policies, deletion controls, data export for compliance, or privacy/GDPR settings for AI-native users; it covers only design, collaboration, export, and MCP/AI features. Missing for 10: any documentation of data retention settings, account/file deletion controls, or privacy compliance mechanisms.

   ⚑ flag

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

   weight 2 · round drawn
   Canva—none0/10

   No evidence pack item addresses telemetry/usage-tracking opt-out settings for AI-native users; documentation covers APIs, SDKs, MCP server, audit logs, and provisioning but nothing about disabling telemetry collection.

   ⚑ flag

   Spline—none0/10

   No evidence in the pack addresses telemetry, analytics, or usage tracking opt-out settings for Spline; the docs cover design features, MCP, exports, and collaboration but nothing about privacy/tracking controls. This is a fair axis for a cloud-based SaaS design tool, so absence of evidence yields 'none'.

   ⚑ flag

1. designerPrototypes can carry advanced behavior — animation between states, conditions or variables, scroll effects — to feel like the real product

   weight 1 · round to Spline
   Canva—none0/10

   The evidence pack covers Canva's developer APIs, SDKs, MCP server, and automation/integration features, but contains no mention of interactive prototyping capabilities such as animated state transitions, conditional logic, variables, or scroll-triggered effects within Canva designs. This is a fair axis for a design tool to be judged on, but nothing in the pack shows Canva delivers it.

   ⚑ flag

   Spline✓fullclaimed8/10

   Spline's docs strongly evidence advanced prototyping behavior: states/events/actions interaction model, variables and real-time API, webhooks for event-driven reactions, animation, multi-scene transitions, physics, and even code/HTML embedding for custom logic — all built to make prototypes feel like real products. Missing for 10: independent hands-on validation of complex behavior chains (scroll-driven effects specifically) and no dedicated 'scroll trigger' doc beyond general events/actions system.

   • [claimed-docs] “Most interactivity in Spline needs no code at all. It's built from **states, events and actions**”
   • [claimed-docs] “The Real-time API feature in Spline allows you to integrate external services, retrieve or send data, and automate responses based on events…”
   • [claimed-docs] “Webhooks in Spline allow you to create real-time, event-driven interactions by listening for and responding to external events.”
   • [claimed-docs] “By using webhooks, your Spline project can automatically trigger actions, update variables, and react to data sent from external sources”
   • [claimed-docs] “You can create multiple scenes within the same Spline file... It's possible to transition between scenes with the Events system by using a S…”
   • [claimed-docs] “On Spline you can create different types of physics simulations, each object can be adjusted to react to gravity and collision forces.”
   • [claimed-docs] “A file can carry one HTML document — markup, CSS and JavaScript — that the editor renders in a sandboxed iframe layered exactly over the 3D …”
   • [claimed-docs] “It enables you to trigger actions, modify object properties, and create dynamic, responsive 3D environments directly from your existing code…”

   proof ↗⚑ flag

2. designerWire screens into clickable prototypes with triggers, transitions, and overlays — no code required

   weight 2 · round to Spline
   Canva—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

   Spline~partialclaimed6/10

   Docs confirm no-code interactivity via states/events/actions (spline-docs-31) and multi-scene transitions via the Events system's Scene Transition action (spline-docs-61), which directly supports triggers and transitions for prototyping. However, 'overlays' are only evidenced via the Code tab feature, which requires writing HTML/CSS/JS (spline-docs-11/24), contradicting the 'no code required' framing for that specific element, and there is no dedicated hands-on or independent account of designers building full clickable multi-screen prototypes end-to-end. Missing for 10: a no-code overlay/layering mechanism, and independent/hands-on evidence of a complete click-through prototype workflow across screens.

   • [claimed-docs] “Most interactivity in Spline needs no code at all. It's built from **states, events and actions**”
   • [claimed-docs] “You can create multiple scenes within the same Spline file... It's possible to transition between scenes with the Events system by using a S…”
   • [claimed-docs] “A file can carry one **HTML document** — markup, CSS and JavaScript — that the editor renders in a sandboxed iframe layered exactly over the…”
   • [claimed-docs] “A 2D canvas for interfaces and animated graphics — vector networks, effects, typography, and real 3D shapes lifted straight out of flat vect…”

   proof ↗⚑ flag

Canvas

1. designerFrames can behave responsively — auto layout, constraints, and grids resize designs the way real UIs flex

   weight 2 · round drawn
   Canva—none0/10

   The evidence pack covers Canva's APIs, SDKs, MCP server, and admin/integration tooling, but contains no mention of responsive frames, auto layout, constraints, or grid-based resizing behavior akin to real UI flex layouts within the design editor itself. Design resizing mentioned (canva-docs-6, canva-docs-16) refers to static output resizing to new fixed dimensions, not dynamic responsive layout behavior.

   ⚑ flag

   Spline—none0/10

   Spline's docs cover 3D modeling, materials, animation, multiplayer, AI generation, and code export, and even mention a 2D canvas (Hana) for interfaces, but nowhere is there mention of auto layout, constraints, or responsive grid systems that would let frames resize like real UI layouts. Missing for 10: any documentation of auto-layout, constraint-based resizing, or grid systems for frames.

   • [claimed-docs] “A 2D canvas for interfaces and animated graphics — vector networks, effects, typography, and real 3D shapes lifted straight out of flat vect…”
   • [claimed-docs] “Hana's **AI Agent** designs for you. Describe a screen and it builds it on the canvas: layout, type, colour, effects, states and interaction…”
   • [claimed-docs] “describe a screen and Hana designs it, then edit it by hand”

   proof ↗⚑ flag

2. designerI get professional vector tools — pen and path editing, boolean operations, precise alignment — for real interface design work

   weight 3 · round to Spline
   Canva—none0/10

   The evidence pack covers Canva's developer APIs, SDKs, MCP server, and automation/integration features, but contains no mention of pen/path tools, boolean operations, or precision alignment controls for vector editing. Canva is a template-driven design tool, so this axis is fair to ask, but nothing in the evidence supports professional vector-editing capability.

   ⚑ flag

   Spline~partialclaimed5/10

   Docs confirm a pen tool for custom 2D shapes (spline-docs-43), boolean operations (Union/Subtract/Intersect) for objects (spline-docs-60), a dedicated 2D canvas with vector networks for interfaces (spline-docs-38), and extrusion turning 2D shapes into 3D (spline-docs-42) — core vector-editing primitives are present. However, there's no documentation of precise alignment/distribution tools, grids, or snapping, and no independent/hands-on evidence validating these tools for serious interface design work versus its primary 3D focus. Missing for 10: alignment/snapping tooling documentation, path-editing depth (e.g., bezier handle controls), independent corroboration of professional UI design usage.

   • [claimed-docs] “The pen tool enables you to create custom 2D shapes in the 3D space. It works in a familiar way to other 2D or UI design tools.”
   • [claimed-docs] “There are 3 different Boolean Operations that you can apply to a set of selected 3D objects: Union ... Subtract ... Intersect”
   • [claimed-docs] “A 2D canvas for interfaces and animated graphics — vector networks, effects, typography, and real 3D shapes lifted straight out of flat vect…”
   • [claimed-docs] “Although 2D objects don't have a depth parameter, they can quickly become 3D using a modeling operation called **Extrusion**.”

   proof ↗⚑ flag

Files

1. designerExport production assets — PNG, SVG, PDF, and more — with per-layer presets, scales, and slices

   weight 2 · round to Canva
   Canva~partialclaimed4/10

   Canva's docs confirm multi-format export (PDF, PNG, JPG, PPTX, MP4) and automatic resizing to new dimensions/aspect ratios, covering the basic 'PNG, SVG, PDF, and more' export need, but there is no evidence of SVG export, per-layer export presets, or slice-based export controls akin to design-tool layer/slice workflows. missing for 10: SVG export support, per-layer export presets, and slice/scale export tooling.

   • [claimed-docs] “Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “Design export:** Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “Automatically create resized versions of designs in new dimensions, formats, or aspect ratios.”
   • [claimed-docs] “Canva Pro and above: Design resizing to new dimensions”

   proof ↗⚑ flag

   Spline—none0/10

   Spline's documented export options are all code/runtime based (Vanilla JS, Three.js, React, Next.js, react-three-fiber, self-hosted ZIP, spline-viewer web component) with no mention of PNG/SVG/PDF raster/vector asset export, per-layer presets, scales, or slices anywhere in the evidence pack.

   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS, Three.js, React, Next.js, react-three-fiber”
   • [claimed-docs] “Spline allows you to export the code of your scene as: **Vanilla JS**... **Three.js**... **React**... **Next.js**... **react-three-fiber**”
   • [claimed-docs] “The Self-Hosted export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a fully se…”
   • [claimed-docs] “allows you to export your entire project, including all assets and the Spline runtime, as a fully self-contained ZIP file”
   • [claimed-docs] “Code & Self-hosted exports * File version history * Team Library”

   proof ↗⚑ flag

2. developerThe file format is open or documented, so my designs stay portable and toolable outside the vendor's editor

   weight 1 · round to Spline
   Canva—none0/10

   Evidence shows only APIs (Connect API, Apps SDK, MCP) and export formats (PDF/PNG/PPTX/etc.) for integration, but no documented open native design file format — Canva's underlying .canva design representation itself remains proprietary and undocumented; users must go through the vendor's own APIs/exports to get data out, not a portable open format usable in other tools without Canva's platform.

   • [claimed-docs] “Design export: Generate exports in multiple formats (PDF, PNG, JPG, PPTX, MP4, and so on) and resize designs for different channels”
   • [claimed-docs] “The Canva Connect APIs let you integrate Canva's capabilities into your own web apps and platforms.”
   • [claimed-docs] “Canva provides an OpenAPI description of the Connect APIs. You can use this with code generation libraries...to generate client SDKs in vari…”
   • [probe] “PROBE openapi: all candidate paths 404 (https://www.canva.dev/openapi.json, https://www.canva.dev/swagger.json, https://www.canva.dev/api/op…”

   proof ↗⚑ flag

   Spline~partialclaimed5/10

   Spline's native .spline file format itself is never documented as open or given a public schema; instead portability is achieved indirectly via export mechanisms — code export to Vanilla JS/Three.js/React/Next.js/react-three-fiber, a self-hosted ZIP with the runtime, and a Code API — which let designs be used outside the editor without exposing/documenting the native format. missing for 10: an open/published spec for the native .spline file format, third-party tooling that reads/writes it directly, independent confirmation that exported code is a full fidelity substitute for the file itself.

   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS, Three.js, React, Next.js, react-three-fiber”
   • [claimed-docs] “The Self-Hosted export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a fully se…”
   • [claimed-docs] “The Code API allows you to seamlessly integrate interactive 3D experiences into your projects, offering greater control and customization.”
   • [claimed-docs] “Spline allows you to export the code of your scene as: **Vanilla JS**... **Three.js**... **React**... **Next.js**... **react-three-fiber**”
   • [claimed-docs] “Spline allows you to export the code of your scene as: Vanilla JS ... Three.js ... React ... Next.js ... react-three-fiber”
   • [claimed-docs] “The **Self-Hosted** export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a full…”

   proof ↗⚑ flag

3. developerSelf-host the full design platform on my own infrastructure under an open-source license

   weight 2 · round drawn
   Canva—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

   Spline—none0/10

   Spline is a cloud-based, closed-source SaaS design tool; the only 'self-hosted' feature documented is exporting a finished 3D scene as a self-contained ZIP for use in your own web app, not self-hosting the design editor/platform itself. There is no evidence of an open-source license, on-prem installer, or self-hostable server for the full editor.

   • [claimed-docs] “The Self-Hosted export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a fully se…”
   • [claimed-docs] “allows you to export your entire project, including all assets and the Spline runtime, as a fully self-contained ZIP file”
   • [claimed-docs] “The **Self-Hosted** export option in Spline allows you to export your entire project, including all assets and the Spline runtime, as a full…”
   • [claimed-docs] “Code & Self-hosted exports * File version history * Team Library”
   • [claimed-docs] “Spline runs in the browser on any platform, with nothing to install and no GPU requirements to check.”
   • [claimed-docs] “It's web-based. Spline runs in the browser on any platform, with nothing to install and no GPU requirements to check.”

   proof ↗⚑ flag

4. designerFull version history lets me name milestones, compare, and restore any earlier state of a file

   weight 2 · round to Spline
   Canva—none0/10

   The evidence pack covers Canva's developer APIs, SDKs, MCP server, and admin/audit features, but contains no mention of file version history, naming milestones, comparing versions, or restoring earlier states of a design.

   ⚑ flag

   Spline~partialclaimed6/10

   Docs confirm a Version History feature that lets users preview all saved versions, track changes, and restore/recover earlier file states, plus manually save new versions ([spline-docs-15],[spline-docs-29],[spline-docs-33]). However, there is no evidence of explicitly naming/labeling milestones or a side-by-side compare view between versions. Missing for 10: explicit naming/labeling of milestone versions, a diff/compare UI between versions, independent hands-on confirmation of restore workflow.

   • [claimed-docs] “Version History allows you to access and preview all the saved versions of your files. This feature helps you track changes, recover previou…”
   • [claimed-docs] “Version History allows you to access and preview all the saved versions of your files.”
   • [claimed-docs] “Code & Self-hosted exports * File version history * Team Library”

   proof ↗⚑ flag