Last updated: 9/15/2026Last verified: 2026-07-20
What is deja-vu?
deja-vu is a local-first session index and MCP recall tool for AI coding workflows. It is designed to search and hand off past conversations from environments such as Claude Code, Codex, Cursor, OpenCode, Grok Build, and other AI coding tools. The project focuses on helping developers reuse prior context across tools without relying on a single hosted workspace.
Key Features
- Indexes and searches historical sessions across multiple AI coding tools
- Provides relevant context recall through an MCP server
- Supports syncing histories over SSH to hand work off between environments
- Includes sensitive content redaction features
- Can optionally add semantic recall using Ollama or LM Studio
- Designed for local-first developer workflows
Best For
Pricing
deja-vu is listed as free. Because the project is hosted on GitHub, users should review the repository for current installation requirements, license details, and any changes to availability.
Pros & Cons
Pros
- Helps unify conversation history across several AI coding environments
- Local-first approach may appeal to developers who want more control over their session data
- MCP support can make prior context available inside compatible AI workflows
- SSH sync is useful for moving histories between machines or environments
- Optional semantic recall can be added with Ollama or LM Studio
- Redaction support helps prepare histories before recall or sharing workflows
Cons
- Requires setup from a GitHub project rather than a polished hosted SaaS interface
- Usefulness depends on whether your AI coding tools and session formats are supported
- Semantic recall requires additional local tooling such as Ollama or LM Studio
- It is focused on AI coding session history, not general project management or documentation
- Redaction behavior should be tested before relying on it for sensitive repositories
Alternatives
Offers developer-focused workflow memory and context retrieval across coding activities, making it relevant for users who want persistent AI-assisted development context.
An AI coding assistant that integrates with IDEs and supports customizable models, suitable for developers who want an extensible coding workflow rather than a dedicated session index.
An AI-native code editor with built-in coding assistance and project context features, useful for developers who prefer an integrated editor experience.
Provides AI coding assistance with codebase context, making it an alternative for teams that need repository-aware help inside development workflows.
An AI coding environment focused on agentic development workflows, suitable for users looking for a broader AI coding workspace.
FAQ
Details
Platform
Features
- Index and search historical sessions across multiple AI coding tools
- Recall relevant context through an MCP server
- Sync histories over SSH and hand work off between environments
- Redact sensitive content and optionally add semantic recall with Ollama or LM Studio
Languages
Known limitations
- Indexing depends on the local history formats and files exposed by each supported coding tool
- Some OpenCode and Cursor indexers require the sqlite3 command-line tool
- Semantic recall is optional and requires a local Ollama or LM Studio model runtime
- Windows builds are available, but macOS and Linux are the more battle-tested paths
