Deja Vu
Deja Vu is an open-source Go command-line tool that indexes local coding-agent session histories and lets users search or recall prior work across 16 documented harnesses, including Claude Code, Codex, OpenCode, Cursor, Gemini CLI, Qwen Code, Kimi Code, Goose, and Copilot CLI. It can expose recall through MCP and inject relevant context at session start where the harness supports hooks.
Coding-agent sessions often repeat investigations because useful context is buried in old transcripts. Deja Vu makes those histories searchable without requiring a hosted memory service, but users still need to understand local-data boundaries, secret redaction, optional sharing or sync, and platform maturity.
The vshulcz/deja-vu repository, documentation, and releases describe install, search, MCP, recall, privacy, redaction, sharing, sync, and platform support. Version 0.17.1 moves more harness guidance into one shared on-demand skill, expands `/deja` command and hook coverage, turns Copilot and Kimi tool activity into work records, and makes the capability matrix explain missing integrations. It also fixes transcript-ID collisions, path-redaction case handling, recall wiring, imported-record indexing, and search ranking, while reporting lower CJK index-build overhead. Community launch discussion can surface practical questions about recovering prior agent work, but it is not independent security or performance proof.
- Search earlier Claude Code, Codex, and OpenCode sessions.
- Recall prior debugging or implementation decisions at session start.
- Expose local agent-history search through MCP.
- Share a deliberately selected session after reviewing redaction and network boundaries.
Deja Vu scans supported local session histories, builds a lexical index, and lets users search prior conversations from the terminal. Its MCP server and session-start integrations can surface relevant earlier work before a coding agent repeats the same investigation. Version 0.17.1 standardizes more guidance as a shared on-demand skill, adds `/deja` command access for OpenCode, Cursor, Roo, and Gemini, wires Grok auto-recall through supported hooks, and converts Copilot and Kimi tool activity into searchable work records.
The repository says indexing and search stay local and include secret redaction. Version 0.15.7 says the recall kill switch and policy now bind every recall path. Sharing, statistics, and sync are explicit features with different data boundaries, so teams should review each command before enabling it and should not equate local search with every optional networked workflow.
Deja Vu documents history discovery across 16 harnesses, but MCP and session-start hook support differ by client. It ships Windows builds while describing macOS and Linux as the battle-tested platforms. Teams should verify session-path coverage, repository exclusion needs, redaction behavior, index storage, sync boundaries, and whether lexical recall is sufficient for their history volume.
Session recall is part of the state and handoff layer around coding agents.
Historical sessions are useful evidence only when their provenance and boundaries remain visible.
Local history, MCP access, redaction, sharing, and sync need explicit permission boundaries.
Deja Vu FAQ
Common questions about Deja Vu.
Does Deja Vu upload coding-agent sessions?+
The repository says indexing and search run locally without network access. Optional sharing, statistics, or sync commands have separate network boundaries, so review the exact command and redaction behavior before using them with sensitive histories.
Which coding agents does Deja Vu support?+
The project documents local histories from Claude Code, Codex, OpenCode, aider, Gemini CLI, Cursor, Antigravity, Grok Build, Qwen Code, Kimi Code, Cline, Roo Code, OpenClaw, Goose, pi, and Copilot CLI. MCP and automatic session-start recall are not available through every harness, so verify the current integration table before installing.
Does Deja Vu work on Windows?+
The project publishes Windows builds, while describing macOS and Linux as the battle-tested platforms. Windows users should verify local session discovery, path handling, MCP startup, and redaction before relying on it for routine recall.