GitBot
Product Hunt Sunny ☀️
GitBot turns repeat AI coding tasks into reusable bots for your repos. It runs locally, is open source, and says there’s no account or telemetry.
Based on reporting by Product Hunt, Sunny ☀️ — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
GitBot is trying to make AI coding agents behave less like one-off prompts and more like tools you can hand around. The pitch is simple: take a job you keep giving Claude Code, Codex, or OpenCode, write the instructions once, set the boundaries, and turn it into a bot that can run in any repo.
That means each run becomes a thread you can return to instead of starting over from scratch. GitBot also leans into reuse. There’s a Library where you can install bots other devs have made, or share your own with a code.
One example in the launch is ShipGuard, which looks at a branch and then decides whether to merge it or block it, with file and line evidence to back the call. That’s a more pointed use case than generic “AI helper” talk. It’s the sort of thing teams actually argue about.
The other detail that matters is where it runs. GitBot runs on your machine, using the logins you already have. It’s open source, with no account and no telemetry, which will sound very good to people who like their tools local and their AI less nosy.
My take — AI-written commentary, not fact-checked reporting
This is the right kind of AI tooling: narrow, local, and not begging for another account. The industry keeps trying to sell “agents” as if the missing piece were bigger ambition, when most developers just want repeatable chores that don’t phone home. Open source plus no telemetry is not a cute extra here; it’s the main event.
Read more about this at: Product Hunt