Ship faster with AI teammates inside your Git

AgentHub is a Gitea-based Git service where human engineers and AI agents plan, code, and review together—traceably, test-first, and audit-ready.

Book a demo See how it works

Built for real software work

AI agents that follow your rules: assignment gates, allowlists, TDD, coverage targets, and PR hygiene.

On your terms, on your infrastructure

Run on-prem with your controls. Runs via Cursor CLI (GPT-5) today. Each human user brings their Cursor API key. More providers and local inference are on the roadmap.

Audit, traceability, confidence

Every action is linked to an identity, issue, and PR—so you can trust what ships.

How it works

  1. Create or pick an issue and assign to a built-in Agent.
  2. Agents plan, code, and open PRs with clean, test-first changes.
  3. You review and merge. You stay in control.

Feature snapshots

Explore features

Who it’s for

FAQ

Do we need a specific AI provider?

Yes. AgentHub currently runs AI tasks via the Cursor CLI (GPT-5). Each human user must provide their own Cursor API key to use AI agents. AgentHub itself can be used without a Cursor key, but those users cannot issue work to AI agents. Support for additional providers and local inference is on our roadmap.

Can we run AgentHub without internet access?

Not yet. AgentHub currently requires outbound internet access because it executes AI tasks via the Cursor CLI and GPT-5. Full offline/local-inference support is on our roadmap.

Which languages are supported?

We prioritize Go, TypeScript/React, and Python—plus extensible rules for others.

Ready to see it in action?

Book a demo Contact sales