Launching November 13, 2029
Orbi

Orbi

GitHub Issues in, tagged releases out.

FreeAI AgentsOpen Source4 impressionslaunches Nov 131 comment

Comments

>log in to comment
  • Larry & Leo[maker]· 15d ago

    Hey everyone 👋 I'm Lawrence, the maker. Most AI coding tools stop at the PR. Orbi keeps going: a second session checks the diff against the Issue's acceptance criteria, then it merges and tags the release. When it gets stuck it labels the Issue ai-blocked and waits for a human. Happy to answer anything about running it unattended.

Orbi is an open-source (AGPL-3.0), self-hosted autonomous coding agent written in Python. Label a GitHub Issue ai-ready and it writes the code in an isolated worktree, opens a PR, has an independent review session check it against the Issue's acceptance criteria, merges only the reviewed head, and cuts a tagged release. GitHub Issues are the only state store: no database, queue, or daemon. A systemd/launchd timer ticks every 5 minutes, so normal operation needs no supervision. When something fails it stops and marks the Issue ai-blocked for a human instead of silently falling back. Every log line and progress comment carries the same run_id, so one grep rebuilds the full timeline. The repo is its own ledger: Orbi's merged PRs, closed Issues and releases were shipped by Orbi. Install: uv tool install orbi-cli (Python ≥ 3.14), or Docker (GHCR / Docker Hub). Don't want to run it yourself? Orbi Cloud runs the same loop: try 3 deliveries on your own repository free, no card.

Orbi is an open-source autonomous AI agent that turns GitHub Issues into reviewed pull requests and tagged releases.

for
Teams that want AI-driven end-to-end issue delivery without leaving code in a third-party service.
pricing
freemium · free trial
license
AGPL-3.0
orbi-build/orbi 195 23Pythonupdated 14 days ago
works withGitHubPythonDockersystemd/launchd

Key features

8 features of Orbi
  • Issue-to-release pipeline — Labels an Issue as ai-ready, writes code in an isolated worktree, reviews, merges and tags a release.
  • Independent review session — A second AI session reviews the PR, can fix failures and re-runs tests before merge.
  • GitHub-native state — All state lives in GitHub (issues, PRs, labels, run_id), no external DB or daemon.
  • Self-hosted or cloud — Run the open-source core on your own machine or use Orbi Cloud with managed runners.
  • Bring-your-own model — Supports any OpenAI-compatible API or locally hosted model for code generation.
  • Automatic recovery — If a run crashes, the next timer tick resumes from the same GitHub record.
  • Token-based cloud pricing — Cloud plans include monthly token allowances; no overage billing.
  • CLI & Docker install — Install via uv tool or Docker image; requires Python 3.14, git, gh, systemd.

Use cases

  • Automatically implement a backlog item labeled ai-ready and ship a tagged release.
  • Run AI-generated code through an independent repair session before merging.
  • Self-hosted CI where code, credentials and test suites never leave the organization.
  • Managed cloud delivery for teams that prefer not to maintain the runner infrastructure.
  • Free$03 merged deliveries free · No card required · Limited to 1 repository
  • SoloUS$24/month (yearly US$290)400M tokens/month · 1 repository · Managed Cloud
  • ProUS$66/month (yearly US$790)1.2B tokens/month · 5 repositories · Managed Cloud
  • Self-hosted$0Open source AGPL-3.0 · Bring your own model · No platform fee

Orbi vs alternatives

OrbiFineSuperAGI CloudPieces CopilotCalljmp
Best forAI-driven issue-to-release automationInteractive AI coding assistanceGeneral AI agent orchestrationAI code suggestionsManaged AI agent backend
PricingFreemiumFreeFreeFreeFree
DevHunt upvotes175436335
LaunchedNov 2029Jan 2023Jan 2023Mar 2024Feb 2026
  • Orbi vs Fine: Fine provides AI agents for building software but focuses on interactive assistance, not end-to-end issue-to-release automation.
  • Orbi vs SuperAGI Cloud: SuperAGI Cloud lets you run autonomous AI agents on cloud infrastructure, but does not integrate tightly with GitHub Issue workflow.
  • Orbi vs Pieces Copilot: Pieces Copilot is an AI coding sidekick that suggests code while you work, not a full delivery pipeline.
  • Orbi vs Calljmp: Calljmp offers managed AI agents as code, lacking the GitHub-native issue-to-release loop Orbi provides.

Orbi FAQ

What does Orbi need to run self-hosted?+

A Linux or macOS machine with Git, a POSIX shell, Python 3.14+, gh CLI and access to an OpenAI-compatible model API.

Can I use my own AI model?+

Yes, self-hosted Orbi works with any OpenAI-compatible API or locally hosted model; you pay the model provider directly.

How is Orbi different from Copilot or Cursor?+

Copilot and Cursor assist while you type; Orbi runs autonomously from an Issue through review, merge and release without human typing.

Who reviews the AI-generated code?+

A separate Orbi session independently reviews, can modify the code, reruns tests and decides whether to merge.

What happens if a delivery crashes mid-run?+

State is stored in GitHub; the next timer tick resumes from the same Issue, PR and run ID, reconstructing the delivery.

Is the cloud service free to try?+

Orbi Cloud offers 3 free merged deliveries with no credit card; after that you subscribe to Solo or Pro plans.

Summarized by DevHunt from orbi.build · Oct 1, 2026. Details may change; check the official site.