Back to directory

deja

deja is a Go binary that gives coding agents shared memory by indexing session history that Claude Code, Codex, Cursor and 30-plus other agents already write to disk, handing the right one back to whichever agent asks, with no model call or capture step.

deja

deja builds a coding-agent memory from the session history already on disk. It indexes the sessions written by Claude Code, Codex, Cursor and more than 30 other supported agents, then hands the relevant prior session back to whichever agent asks, without running a model, adding a capture step, or running a server.

The project publishes a benchmark comparing its accuracy, cost and speed on the LongMemEval-S and LoCoMo benchmarks against other coding-agent memory tools it names as agentmemory and MemPalace, reporting the highest recall at rank five, roughly half the token cost, and the fastest indexing and search times of the tools it measured. It installs with a shell script, Homebrew, a Go install, or an npx one-off run, and wires MCP-based recall into every supported agent it finds on a machine, including automatic recall at session start where an agent supports it.

More in Agent memory

See all 450
  • A long technical guide to AI agent memory: why stateless LLMs need it, the different memory architectures from vector stores to knowledge graphs, and how Cognee implements one as an open-source memory layer.

  • A local, file-based long-term memory layer for a personal AI that never summarizes conversations, keeping every line with a timestamp and searching by time first and words second, rather than retrieving only similar-seeming chunks.

  • NEOTH

    deepwiki.com

    A local-first personal AI daemon in Rust with five memory tiers, consent-gated tools, a WASM plugin sandbox, multi-provider model routing and signed audit logs, aimed at people who want both a daily assistant and a serious coding helper.