Point a coding agent at research and keep every experiment
Made by alphaXiv
View profile on GitHub (opens in a new tab)A local workspace that turns coding agents such as Claude Code, Codex, OpenCode or Cursor into research assistants: reviewing literature, proposing a hypothesis, changing code, launching an experiment and deciding what to try next.
Everything runs on your own machine at a local address with a local database, so creating a project or launching a run does not publish your code. An experiment tree keeps the lineage when several agents explore different directions at once, and the same committed snapshot can be run locally, over SSH, or on a cluster or cloud GPU service.
Can I use this?
You'll need macOS 11+ or Linux, plus a coding agent · Setup needed
Worth knowing One warning in the project’s own documentation is worth reading twice: when you run it on a remote machine the service listens on that machine’s loopback address with no login of its own, so anyone else with an account on that host can reach it. Windows support is described as beta and needs Git for Windows. Official release builds send coarse usage events tied to a random install id, which can be switched off; the project says they exclude code, prompts, file contents, paths and repository names. MIT licence. BuildTube has not run or verified this software.
Why it's here
Not trending this week
Last seen on the GitHub trending list on 20 September 2026. It stays on BuildTube so you can still find it.
- 6,670 people have starred it on GitHub.
Numbers checked on 6 October 2026; not refreshed since.
Behind it
See the code on GitHubBuildTube has not run or verified this project. Everything above is written from what the creator published.
Made this?