Preview · GA 2026–27
PreviewGluecron builds and ships itself

The git host that
does the work.

Every push is gated, every PR reviewed by Claude in under 30 seconds, every red check repaired before you notice. You stay in control.

claude mcp add --transport http gluecron https://gluecron.com/mcp
acme/apifix/token-refresh → main · 4b1e9c2Example push
AI review · every PR
<30s

Line-level comments, risk flags and a verdict your merge gate can require.

Spec → PR
74s to a draft PR

Label an issue ai:build. Claude branches, writes the change and opens a draft PR against your gates — spec to PR in under 90 seconds.

Sleep mode · overnight
3 PRs merged while you slept

Autopilot merges ready PRs, builds labelled issues and repairs failed gates — and tells you in the morning.

Built for both

GitHub was built for humans typing. This one was built for both.

Claude doesn't get a chat box bolted onto a git host. It gets the git host — a bot account, your permissions, and its commits in your history like any other contributor.

MCP-native, one line

117 MCP tools · Claude Code, Claude Desktop, Cursor & any MCP client · all install options →

claude › open a PR for issue #42
→ gluecron_create_branch feat/export
→ gluecron_write_file src/export.ts
→ gluecron_create_pr #43 · gates armed

Secure at push

Secret scanning, CVE alerts and push policy run at the HTTP layer, before the branch moves — enforced, not advice.

✗ push rejected — AWS key
config/prod.env:3 · AKIA••••7Q2X
branch unchanged · 0.4s

Yours to host

One Bun server, Postgres and plain git repositories. Self-hosting is free, with no licence key and no telemetry.

$ docker compose up -d
✓ gluecron · :3000 · git over HTTPS
✓ first sign-up becomes the admin
The platform

A complete git host underneath.

The AI only matters if the platform under it is real. Issues and pull requests, CI, orgs and teams, packages and audit logs — all first-party, none of it an add-on.

Code hosting

  • Git over HTTPS (Smart HTTP)
  • Forks · stars · topics
  • Releases · protected tags
  • Push policy rulesets

Collaboration

  • Issues · PRs · inline review
  • Draft PRs · merge queues
  • Projects · discussions · wikis
  • Organizations · teams · roles

CI / CD

  • YAML-native workflow runner
  • Branch protection · required checks
  • Preview deploys per PR
  • Auto-repair on gate failure

Security

  • Secret scanning at push
  • Dependency CVE alerts
  • 2FA · passkeys · SAML SSO
  • Audit log across the platform
Packages, Pages, DORA metrics, GraphQL and the rest — see every feature →
Pricing

Free to start. Honest at scale.

Self-hosting is free forever. Hosted plans price the AI calls, and you own your data either way.

Free
Free forever
  • 10 repos
  • 100,000 AI tokens / month
  • 1 GB storage
  • 10 GB bandwidth / month
ProMost popular
$9.00/mo
  • 200 repos (public + private)
  • 1,000,000 AI tokens / month
  • 10 GB storage
  • 100 GB bandwidth / month
Team
$24.00/mo
  • 1,000 repos (public + private)
  • 5,000,000 AI tokens / month
  • 50 GB storage
  • 500 GB bandwidth / month

Full pricing and plan details →Preview pricing — final plans are published at general availability.

Invite-only during build-out

Put a git host to work on your next PR.

Tell us what you build and we will send an invite code. Already have one? Create your account in a minute.

Have an invite code? Create your account →