# Store and search agent memory

Save embeddings or notes and retrieve the most relevant ones later.

Also asked as: vector database, semantic search, agent memory.

## Websites for this task

None in the directory yet.

MCP: `find_sites({"task":"store-agent-memory"})`.

## APIs and MCP servers in the catalog

- [Chroma](https://agentready-tools.com/tools/chroma): Open-source and hosted vector database for AI applications, with official APIs and usage-based Chroma Cloud pricing.
- [Pinecone](https://agentready-tools.com/tools/pinecone) (MCP): Managed vector database and knowledge platform for search, retrieval, and AI applications, with official APIs and an official MCP server.
- [turbopuffer](https://agentready-tools.com/tools/turbopuffer): Vector and full-text search database with official APIs and usage-based cloud pricing.
- [Weaviate](https://agentready-tools.com/tools/weaviate): AI database with official REST and GraphQL APIs, SDKs, and a managed cloud.
- [Letta](https://agentready-tools.com/tools/letta): Platform for stateful AI agents, with official docs, APIs, and credit-based model and sandbox usage.
- [Mem0](https://agentready-tools.com/tools/mem0) (MCP): Memory layer for AI agents and apps, with an official API and a hosted MCP server at mcp.mem0.ai/mcp.
- [Qdrant](https://agentready-tools.com/tools/qdrant): Vector search engine and managed cloud for similarity search, available as open source and via official APIs.
- [Zep](https://agentready-tools.com/tools/zep) (MCP): Memory and context platform for AI agents, with an official API and Context MCP server seats on paid plans.
- [Upstash](https://agentready-tools.com/tools/upstash) (MCP): Serverless Redis, Vector, and related data APIs, plus an official Upstash Redis MCP server, authenticated with API tokens.

MCP: `search_tools({"query":"vector database"})`. Categories: Vector databases and memory, Managed databases and backend platforms.

Updated 2026-11-22. All tasks: https://agentready-tools.com/task.md
