Pipecat Memory
https://github.com/supermemoryai/pipecat-memory- Category
- AI Tools
- Rank
- No. 1649Tools index
Previous survey · No. 1654 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- supermemoryai
- GitHub
- 22 stars
- Date
About
Adds persistent memory to Pipecat voice AI agents, so conversational bots can remember context across calls.
What it does
Pipecat Memory sits between conversation context collection and the language model. It can supply user facts, search prior material, or combine both behaviors. The included demo shows the result through a browser voice client, transcripts, memory listings, profile data, and a memory graph.
Why it's ranked here
The integration has a clear architectural boundary and a runnable full-stack example. Its strongest case is practical: developers can see where memory enters the voice pipeline, how users and sessions are separated, and how stored material becomes searchable context. The narrow automated coverage keeps the verdict measured.
What's good
The three memory modes make retrieval intent explicit. User and session identifiers separate durable identity from individual conversations. The demo also exposes memories, profiles, semantic search results, and graph data, which helps developers inspect what the assistant may recall instead of treating memory as an invisible subsystem.
Tradeoffs
The repository shows only one automated test, covering a worker response for an unknown route. The demo depends on Supermemory and Gemini credentials, while the quick start names OpenAI and Supermemory keys. Listing memories also fetches each document individually in sequence, and several API responses use loose, fallback-heavy data shapes.
How to use it well
Use it when building a Pipecat voice assistant that needs user-specific recall and you want an inspectable reference implementation. Start with profile mode for known facts, query mode for retrieval, or full mode for both. It does not replace speech transport, voice activity detection, the language model, text-to-speech, authentication, or deployment infrastructure.
Technical notes+
backend/server.py builds a FastAPI WebSocket voice pipeline with SupermemoryPipecatService placed after LLMContextAggregatorPair and before GeminiLiveLLMService; it keys memory by user_id and session_id, using full mode, a search limit of 10, and a threshold of 0.1. src/server.ts is a Cloudflare Worker proxy for Supermemory v3 memory, profile, search, document, and health endpoints. src/app.tsx uses the Pipecat JavaScript client, WebSocket transport, React state, browser local storage, and @supermemory/memory-graph. tests/index.test.ts contains one Vitest case asserting a 404 response. package.json defines Bun, Vite, Wrangler, Biome, and Vitest workflows; README.md documents pip install supermemory-pipecat.
Observed
- License
- MIT License
- Primary languages
- TypeScript for the browser and Cloudflare Worker, plus Python for the voice backend
- Package installation
- Python library installed with pip as supermemory-pipecat
- Interfaces
- Pipecat pipeline library, FastAPI HTTP and WebSocket backend, browser client, and Cloudflare Worker HTTP API
- Frontend platform
- React and Vite with Pipecat WebSocket client packages
- Deployment surface
- Cloudflare Worker frontend and proxy, with a separate Python voice backend
- Test structure
- One Vitest worker test is shown, covering the unknown-route 404 response
Read from README.md, package.json, src/app.tsx, src/server.ts, src/client.tsx, env.d.ts, auth-schema.ts, vite.config.ts, vitest.config.ts, drizzle.config.ts, worker-configuration.d.ts, backend/server.py, tests/index.test.ts, LICENSE, bun.lock.
What it can do
Store conversation context across multiple calls
Voice AI agent conversations → Persistent conversation memory
Retrieve previous conversation history
User identifier or session data → Historical conversation context
Maintain user preferences and information
User interaction data and preferences → Stored user profile data
Enable contextual responses based on past interactions
Current user input and stored conversation history → Context-aware AI responses
Track conversation threads and topics over time
Multi-session conversation data → Organized conversation threads and topic history
Tags
Tech Stack
Media
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.