ai-memory
SHA-256ai-memory provides long-term memory for AI coding agents using a git-backed markdown wiki. It captures lifecycle events from Claude Code, Codex, Cursor, Gemini CLI, and others, generating bounded handoffs so you can switch agents mid-task without re-explaining context. No vector database needed — just plain files you can grep and sync.
Smart Download
Download Download Version
v1.28.0 · 11.4 MB
Persistent, grep-able memory for AI coding agents — switch between Claude, Codex, Cursor, and more without re-explaining your project.
Core Features
- Cross-agent memory: supports Claude Code, Codex, Cursor, Gemini CLI, OpenCode, and a dozen more; auto-generates handoff summaries at session end
- Plain Markdown + git storage: memory is grep-able, opens in Obsidian, backs up with rsync — no database or extra service required
- Zero-friction automatic capture: lifecycle hooks record prompts, tool calls, and session boundaries without manual note-writing
- Optional managed workstreams: `ai-memory run` launches agents and resumes the same logical workstream across different tools
What It Can't Do
- •Not all agents have a true session-end hook (e.g., Codex, Kiro CLI). You must manually run `ai-memory finalize-session` at the end of your work to generate the final summary
- •Native Windows support is experimental; using the Linux binary inside WSL2 is recommended
- •Captured observations are sanitized and not a full native transcript — some detail may be lost
- •MCP config and hook installation may require manual adjustments for certain agents; refer to the specific docs
Use Cases
- Developers switching between multiple AI coding agents (e.g., Claude Code and Codex) who want to preserve project context and unfinished tasks
- Teams aiming to turn AI session knowledge into searchable, version-controlled Markdown documentation for long-term accumulation and handoff
Detailed Introduction
ai-memory is an open-source (MIT) Rust tool that solves a common pain point: AI coding agents forget everything when a session ends. It installs as MCP config plus lifecycle hooks across a wide range of agents (Claude Code, Codex, Command Code, Devin CLI, OpenCode, Cursor, Gemini CLI, and others). During a session, it captures sanitized prompts, tool calls, and session boundaries, then compiles them into a coherent summary for the next agent. The result is a portable, grep-able markdown wiki stored in a local git repo — no vector database to maintain, no manual note-writing. It also offers an optional `ai-memory run` command that launches agents inside a managed workstream for higher-fidelity resume across different harnesses. Compared to agent-specific memory plugins or heavy RAG setups, ai-memory is lightweight, vendor-neutral, and designed for seamless handoffs between competing AI coding tools.
Tags
Getting Started
Download installer
Click the button above to download the installer for your system
Install the software
Install the appropriate package for your distro (dpkg / rpm / AppImage)
Download the archive for your platform (Linux/macOS/Windows) from GitHub Releases, extract it, and add the binary to your PATH
Run `ai-memory install-mcp --client <your-agent>` to configure MCP
Run `ai-memory install-hooks --agent <your-agent>` to install lifecycle hooks
- Download the archive for your platform (Linux/macOS/Windows) from GitHub Releases, extract it, and add the binary to your PATH
- Run `ai-memory install-mcp --client <your-agent>` to configure MCP
- Run `ai-memory install-hooks --agent <your-agent>` to install lifecycle hooks
SHA-256 checksum verified
Checksum extracted from GitHub official Release page
SHA256 Checksum
9eed6fb25cc7a057c8e622eb42b83bfcd366858cfb4fd59847a4a4a9ca87a838This checksum is extracted from the GitHub Release page. Verify file integrity after download.
All SHA-256 checksums on this platform are extracted from the project's official GitHub Release page, without any modification. You can independently verify them on the GitHub Releases page.
Open Source Transparency
View GitHub SourceUninstall Info
Run `ai-memory uninstall-hooks --agent <your-agent>` to remove hooks, `ai-memory uninstall-mcp --client <your-agent>` to delete MCP config, then delete the downloaded binary and the `~/.ai-memory` directory.
No Extra Dependencies
Ready to use after download. No additional runtime required.
Similar Projects
deno
Deno is a modern runtime for JavaScript and TypeScript with secure defaults, built-in TypeScript support, and a great developer experience.
Visual Studio Code (Code - OSS)
A free, open-source code editor by Microsoft with rich extensions, debugging, and Git integration. The most popular developer tool worldwide.
opencode
OpenCode is a free, open-source AI coding agent that helps you write, refactor, and understand code. It offers two built-in modes: a full-access 'build' agent for active coding and a read-only 'plan' agent for safe exploration.