Open-source desktop app to run, serve and fine-tune text, image, video and audio models entirely on your own machine.
Free
Summary
Open-source, repo-native project memory for coding agents, stored as plain Markdown.
Screenshots
Description
MemoryCustodian gives coding agents like Claude Code, Codex, and Gemini durable project memory without a hosted service or ever-growing prompt files. Decisions, constraints, rejected approaches, and other project context are stored as plain Markdown files inside docs/memory/ in the repository itself, so they can be reviewed, diffed, versioned, and deleted exactly like code.
A manifest-based routing system loads only the memory relevant to the current task instead of dumping the entire history into every prompt, keeping context windows lean as a project grows. The core protocol is built around a small set of files, manifest.md, brief.md, decisions.md, constraints.md, do-not-use.md, and inbox.md, giving teams a consistent, human-readable place to capture why a codebase looks the way it does instead of losing that reasoning in old chat transcripts.
MemoryCustodian is MIT-licensed, works offline with Python's standard library only, and installs directly into Claude Code, Codex, and Gemini via a simple install script, making it a lightweight fit for teams frustrated that every new agent session forgets the context from the last one.
Reviews
Similar App Suggestions
Spotify's vendor-neutral workspace for running dozens of Claude Code, Codex and Gemini CLI sessions in parallel with shared context.
Free
A coding agent tuned for latency: routes easy work to fast models, pulls only the code it needs, and fans out tool calls in parallel.
Free
Run agent evals and private benchmarks in sandboxed cloud environments — compare Claude Code, Codex, Cursor and Copilot on the same real tasks.
Freemium
An open-source proxy that sits between your coding agent and the model, compressing tool schemas, file reads and stale history to cut token bills.
Freemium