Provides a searchable conversation memory for AI agents, enabling them to recall past discussions and context.
Sponsored
Claw Recall gives AI agents a persistent, searchable memory beyond their immediate context window. It indexes all agent conversations, including OpenClaw and Claude Code sessions, into a local SQLite database, supporting both keyword and semantic search. This tool is crucial for recovering context after compaction, enabling cross-agent knowledge sharing, and allowing agents to access historical data from various sources like Gmail, Google Drive, and Slack, operating at a very low cost.
Key Features
01Indexes OpenClaw and Claude Code agent conversations into a searchable SQLite database
02Supports both keyword (FTS5) and semantic search (via OpenAI embeddings)
03Recovers agent context after compaction or restarts using `browse_recent`
04Captures and indexes external sources including Gmail, Google Drive, and Slack
05Provides access via CLI, REST API, MCP (stdio + SSE), and a built-in web UI
060 GitHub stars
Use Cases
01Allow agents to search and retrieve information from indexed external sources like emails and documents
02Enable cross-agent context sharing, allowing agents to search past discussions of other agents
03Restore an AI agent's context and conversation history after memory compaction
