Letta

Delegate tasks to a built-in AI assistant inside the product G

Agentic features

Drive the product through a documented public API G

Agent access

Plug MCP servers into this product so it can use their tools G

Agent access

Connect an agent via an official MCP server G

Agent access

Build against official SDKs G

Agent access

Operate the product with natural-language commands G

Agentic features

Point an agent at llms.txt or agent-oriented docs G

Agent access

Run the product headlessly / in CI for automation G

Agent access

Set up automations that run autonomously in the background G

Agentic features

Use an official CLI G

Agent access

Get AI-generated insights and suggestions from my data inside the product G

Agentic features

Issue scoped/least-privilege API credentials for an agent G

Agent access

Download a machine-readable API spec (OpenAPI or equivalent) G

Api quality

Explore an interactive API reference with runnable examples G

Api quality

Rely on versioned APIs with a documented deprecation policy G

Api quality

Subscribe to events via webhooks G

Agent access

Test against a sandbox environment without touching production data G

Api quality

Get summaries of past sessions or threads so an agent can pick up where the last conversation left off C

Context assembly

My agent can manage its own memory mid-conversation — adding, searching, updating, and deleting memories through tools or API calls it invokes itself C

Agent memory

Self-host the core product G

Add memories from conversations and retrieve them later with semantic search, so context persists across sessions C

Core memory

Define rules that trigger actions automatically on events G

Self-host the memory layer from open-source code (e.g. via Docker) on infrastructure I control C

Self host

Export all of my data in open formats and leave G

Scope memories per user, agent, or application so one tenant's memories never leak into another's retrieval C

Isolation

Store memories as a knowledge graph of entities and relationships so multi-hop and entity-centric questions are answerable C

Knowledge graph

Prevent my data from being used to train AI models G

The memory layer decides for itself what is worth remembering — extracting salient facts from raw conversation and consolidating them in the background C

Agent memory

Read the product's source under an open license G

Rely on the memory layer to update, supersede, or merge memories when new information contradicts what was stored C

Core memory

Build against official SDKs in at least Python and TypeScript with equivalent memory APIs C

Sdks

Control data retention and deletion G

Do everything through the API that I can do in the UI G

Export memories in a machine-readable format so the memory store is portable and not a lock-in trap C

Portability

Ingest documents, JSON, and business data into memory — not just chat transcripts C

Ingestion

See published pricing with a free tier and per-unit rates so I can project memory costs before committing G

Pricing

Choose where my data is stored (region/residency) G

Perform bulk operations across many items at once G

Steer retrieval with metadata filters, keyword/hybrid search modes, or reranking instead of accepting a single fixed similarity search C

Retrieval controls

Govern who and what can read or write memory with roles, policies, or access-control lists, and audit that access C

Governance

Connect off-the-shelf assistants (Claude, ChatGPT, Cursor) to the same memory so every tool I use shares what it knows about me C

Agent memory

Delete a user's memories on demand — single memory, per-entity, or full erasure — to satisfy privacy requirements C

Forgetting

Make memories expire or decay — via TTL, expiration dates, or recency weighting — so stale facts stop surfacing C

Forgetting

Retrieve a token-budgeted, prompt-ready context block assembled from relevant memories in one call C

Context assembly

See published memory-quality benchmark results (e.g. LongMemEval, LoCoMo) backing the product's recall-accuracy claims C

Benchmarks

Drop the memory layer into agent frameworks like LangChain, LangGraph, CrewAI, or the Vercel AI SDK via documented first-party integrations C

Frameworks

Ingest at scale with async or batch processing and check the status of background memory operations C

Scale

Opt out of telemetry and usage tracking G

Schedule recurring jobs or workflows G

See documented retrieval-latency targets or measured numbers (e.g. p50/p95) backing the product's speed claims C

Latency

Track when facts became valid or invalid (temporal reasoning) so the memory distinguishes current from outdated information C

Knowledge graph

Share selected memory across multiple agents or users (team or group memory) while keeping private memory private C

Sharing

Run the memory layer fully locally — embedded in-process or against local models — without any cloud dependency C

Self host

Version, review, and roll back my automations G

Customize the memory schema — entity types, edge types, or ontology — to match my domain C

Schema customization

Store images, PDFs, or other files as memory inputs and recall information from them later C

Ingestion

Wire memory into real-time voice pipelines (e.g. LiveKit, Pipecat, ElevenLabs) with documented integrations fast enough for live conversation C

Frameworks