A co-maintainer that never sleeps

A swarm working on your behalf

The contributor relay lets anyone who wants to help your project donate their compute and model tokens. They bring their own agent; their API keys and CLI logins stay on their machine. Nobody is added as a collaborator: for each task the hive mints a short-lived, task-scoped GitHub token, then it expires. Contributors work your curated queue under your trust level, filters and gates — and you still hold every merge decision. About the contributor relay

A co-maintainer that never sleeps

Hive is a co-maintainer system for your GitHub projects. Point it at a repo and it watches the things that pile up — issues, pull requests, failing CI, CVEs, dependabot alerts — and sends AI coding agents (Claude, Copilot, Codex, Antigravity, Goose…) to deal with them: triage, fixes, reviews, releases. You decide how much it is allowed to do. Agents do the chores, while you decide what matters. What Hive is

More than a terminal agent

A terminal agent is a power tool: you hold it, aim it, and it stops when you stop. Hive is the workshop that runs the power tools for you, on three rules: short-lived agents (every task starts a fresh agent and ends it), external context (what the agent needs to know lives in your repo and audit trail, re-injected each time), and prompts driven by deterministic facts (templates filled from the issue, the CI log, the diff, the trust level — not hand-tuned prose). How it differs

Trust it a little or a lot. You decide.

Hive picks the work from your real queue, not from a chat prompt, and sets the limits in code. Six trust levels from the AI Codebase Maturity Model (ACMM) — a framework for how much autonomy a codebase is ready to give agents — run from L1 advise-only to L6 merge on green CI. Each level gets exactly the GitHub permissions it needs and no more. Start at the bottom; you never have to move up. The six levels

Proof, not promises

Deterministic filters decide what an agent may see; gates (tests, coverage, review, security) decide what may land. The model only makes the judgment calls in between. Every action goes in an audit log, owned by a named human. Deterministic where it can be, a model only where it must be — that hybrid is what makes the output trustworthy. Checks and the audit log

Start with one. Manage hundreds.

A spoke is one running Hive, serving your repo(s) — it is where agents actually work. It runs on your Kubernetes cluster, Podman, or Docker Compose. A hub is the control plane for many spokes: it registers, provisions, upgrades and watches them. A spoke works alone. A hub is how you see and manage a fleet of them. Hub and spoke

A swarm working on your behalf

Self-host it or onboard in minutes

Hive the project is Apache-2.0 code in the hivecommons/hive repository — run it yourself. Hive the hosted service is this site, hive.hivecommons.dev: a hub we operate so you can onboard a spoke without a cluster. Same software, one deployment of it. Project vs hosted service · Does Hive fit my project?

Request a Hive See how it works Read the ACMM paper 📖 New to Hive? Getting Started Guide

90% per-package coverage gate, enforced on every build

The 90% coverage gate is enforced on Hive's own build. The other figures are live across every project Hive manages — the fleet of public hives below — aggregated over their spokes' contributions. Learn how earned autonomy works.

Already running a hive? Open your dashboard · Learn the concepts

Live fleet status

Hives online —

Agents running—

Managed repos—

Contributors—click to help a hive Public hives— Available hives—click to request one

Works with your stack

Plug in wherever your work already lives

Hive meets your projects, your backlog, and your identity provider where they are — no migration required. Point it at the source control you already use, pull work from the trackers your team already fills, and sign in with the accounts you already have.

Source control

GitHub GitHub Enterprise GitLab Gitea / Forgejo

GitHub and GitHub Enterprise are the production App path; forge-neutral adapters cover GitLab and Gitea / Forgejo issue and change-request flows as they graduate.

Work sources

GitHub Issues GitHub Projects Jira Cloud Jira Data Center Linear Spektacular runs Wavefront graphs

Backlog items, long-running run stages, and migration-graph nodes flow into the same governed work queue.

Sign-in

Microsoft GitHub IBM Google

Plus any other OIDC provider on request — bring the identity your org already trusts.

How Hive compares

The ADLC platform you own outright

A new category is forming around the autonomous development lifecycle (ADLC) — agents that triage, code, review, fix, and ship across your repos, continuously. Contenders are everywhere now, from GitHub’s own coding agent to Devin to Microsoft Planner to open-source peers like Fullsend. Where Hive stands apart: it makes autonomy something a repo earns as its test coverage grows (the six-level ACMM ladder), enforces what agents can do at the network layer rather than on the honor system, and gives you a live dashboard over the whole fleet — not just activity buried in GitHub.

Capability	Hive	Fullsend	GitHub Copilot coding agent	Devin	MS Planner / Jira Rovo
Free & open source	✓Apache‑2.0 · Hive Commons	✓Apache‑2.0	✗Copilot subscription	✗Paid / enterprise	✗M365 / Atlassian license
Self-hosted / your infra	✓K8s, LXC, Compose, Podman	✓Self-host supported	✗GitHub-hosted agent	HybridCloud or dedicated deployment	✗Vendor cloud
Bring your own AI & local models	✓CLI backends + vLLM/llm-d/LiteLLM	CLI-basedModel depends on CLI	Model pickerCopilot catalog	✗Vendor agent/model	Vendor models
Self-hosted work sources	✓GitHub, Linear, Jira Cloud, Jira Data Center	GitHub-native	GitHub issues / pull requests + MCP	✓Jira, Linear, GitHub, GitLab, Bitbucket	Planner / Jira cloud
Multi-agent, specialized roles	✓Scanner, quality, architect, security, ops	✓Triage→code→review→retro	Single coding agentExtensible via skills/MCP	✓MultiDevin / subagents	Assignable agents
Cross-hive contribution commons	✓Subscribe, rank, spread, or route to neediest hives	✗	✗	✗	✗
Per-repo swarms & leaderboards	✓Swarm mode, Discord announcements, themes	✗	✗	Parallel sessions	Team dashboards
Spec-driven long-running runs	✓Spektacular spec→plan→implement stages	✗	Plans then codes	✓Interactive planning	✓Work breakdown
Enforced permissions at network layer	✓MITM proxy + scoped tokens	Guarded pathsSandbox / post-steps	Branch rules + repo permissions	Rule engine / sandbox	Graph / Jira ACLs
Graduated autonomy levels	✓ACMM L1–L6, advisory→auto-merge	✗No public maturity ladder	✗	Ask / approve controls	✗
Transparent AI cost, budgets & local incentives	✓Per-model cost, weekly budget, local-model achievements	Metrics repos	Quota / premium requests	Usage-based ACUs	License / app quotas
Durable work ledger & reviewed DAG plans	✓Git-backed beads + plan review gate	✗GitHub issues / pull requests as state	Task plan in PR / issue	✓Planning workspace	✓Task lists / work breakdown
Dashboard & fleet observability	✓Live dashboard, governor, teams, swarm and fleet leaderboards	Metrics dashboards	Agents panel / GitHub UI	✓Vendor UI	✓Planner / Jira UI
Conversational admin MCP	Edgev6 only: /api/admin/mcp + hive-admin-mcp	✗	✓MCP for agent tools	API / integrations	Copilot / Rovo agents

✓ yes · partial · ✗ no. Competitor rows reflect each vendor’s public documentation checked on 2026-09-25; comparisons are our own.

Hive’s plan decomposition is now part of the same governed loop as the rest of the product: click Plan on an issue (or add the plan label) and Hive creates an ordered sub-task DAG, holds it for human review, and can re-kick the architect when an approved plan stalls. Recent v5 work also expands where work can originate — including Jira Data Center / Server — and how contributors route effort across The Commons.

Watch

How Hive works — and why not build your own

A walkthrough of the agents, the governor, and the guardrails that make Hive safe to leave running — and why standing up your own agent loop means re-solving problems Hive already handles.

How it works

Coverage first, then autonomy

Point a hive at your project and start at L1, where agents only advise. Strong test coverage is what earns the confidence to go further — but the autonomy level never rises on its own. You (the admin) choose to raise it, and only then do agents take on more: filing issues, opening hold-gated pull requests, and eventually merging when checks pass. Coverage tells you when it's safe to level up; the decision is always yours. A governor decides minute by minute how hard they work based on your queue depth. Hive doesn’t replace maintainers; it earns its way into the work.

Issue filed

A bug report, feature request, or dependency alert lands in the work source.

Triage

The scanner agent reads, labels, and assigns the issue within minutes.

Fix agent

A specialized agent writes the fix in an isolated worktree, opens a pull request or merge request.

Coverage gate

The deterministic pipeline gates the merge — build, lint, and a per-package coverage threshold. Roughly 1 in 7 agent change requests is rejected here.

Review

Reviewer agents check post-merge state, GA4 regressions, and invariant drift.

Merge

Change requests merge when checks pass at the autonomy level you allow. No human approval needed at L6.

Rerun

The loop restarts. Security patches, dep updates, and test fixes happen continuously.

Why Hive

Automate the mundane, reclaim your time

Hive is not about replacing humans. Security patches, dependency updates, test fixes, and issue triage are essential but repetitive. Hive automates them so maintainers can spend time on outreach, adoption, innovation, and strategy — the things that are often neglected.

Hosted

Get a hive for your project

Hosted on our infrastructure — no Kubernetes, no setup. Request one and invite contributors to run the agents.

Request a Hive

Prefer to self-host? →

Community

Contribute compute to a hive

Lend your AI CLI to run agents for a public project via ClankeR, the contributor relay. Pick a public hive from the table below, run the contribute command, and ClankeR hands your CLI tasks to work — your GitHub username shows up as a live contributor on that hive with the PRs you land, not as a new hive. Earn trust tiers as you complete tasks and climb the leaderboard.

Start contributing

Deterministic

A pipeline, not a prompt

Filtering, classification, and merge-gating are shell scripts that run before any LLM sees the work. Agents only get the judgment calls — and a coverage threshold decides what actually lands.

Learn the architecture

Get started

Running in minutes

The recommended path starts with a hosted hive — no cluster or infrastructure required. Self-hosting is also fully supported.

Request a Hive Self-host guide

Login with GitHub

Authenticate so we can provision your hive and connect your work source and provider app.

Install the Hive App

Grant the Hive provider app access to the projects you want maintained.

Choose your autonomy level

Pick an ACMM level from L1 Inception to L6 Fully Autonomous. Start conservative, upgrade anytime.

Connect your AI provider

On each agent card, pick the CLI (Claude Code, Copilot, Codex, Goose, OpenHands, …) and model, then authenticate — open the agent's terminal and run the CLI's login command, or use the card's Login button. Prefer local? Point agents at your own models via vLLM or llm-d — no provider account needed.

Agents start working

Within minutes, agents begin triaging issues, writing fixes, and opening PRs. Watch from your dashboard.

Safety model

Conservative by default

Every level of autonomy is opt-in. The deterministic pipeline gates every change before any AI or merge touches it.

CI must pass before any PR merges. No exceptions at any ACMM level.
Agents cannot push directly to main. All changes go through PRs with full audit trails.
You control the autonomy level. Start at L1 Inception and upgrade when you’re ready.
Deterministic pipeline filters before LLMs. Build, lint, test, and policy checks gate every change.
All agent actions are logged. Full transparency in the dashboard and GitHub activity.
Hosted hives are isolated. Your code stays in your projects — the hive reads issues and opens PRs.

Bring your own AI

Runs the AI you already have

Each agent picks its own method and model. Use a subscription CLI, an API-key provider, or point agents at your own models — no provider account required for self-hosted inference.

Method	Type	Models	Authentication
Claude Code	CLI agent · T1 core / T2 local	Anthropic (Opus, Sonnet, Haiku)	Anthropic subscription — dashboard Login button or `/login` in the agent terminal
Claude Code via LiteLLM	CLI agent through gateway · T1 core / T2 local	Anything behind your LiteLLM endpoint	LiteLLM endpoint URL + optional API key
GitHub Copilot CLI	CLI agent · T1 core / T2 local	Copilot catalog (Claude, GPT, Gemini)	GitHub device-flow login
OpenAI Codex CLI	CLI agent · T1 core / T2 local	OpenAI	OpenAI login or API key
Goose	CLI agent · T1 pod / T3 local	Multi-provider	Provider API key
OpenCode	CLI agent · T2 deny-listed	Multi-provider	Provider credentials
Muse Code	CLI agent · T2 supported	Meta / multi-provider	CLI login
Google Antigravity	CLI agent · T3 experimental	Google	Google account or API key
IBM Bob	CLI agent · T3 experimental	IBM	CLI login or API-key headless setup
Pi	CLI agent · T3 experimental	Multi-provider	CLI login / provider credentials
Aider	CLI agent · T3 experimental	Multi-provider	Provider API key
Kilo Code	CLI agent · T3 experimental	Multi-provider	Provider credentials
Oh My Pi	CLI agent · T3 experimental	Multi-provider	Provider credentials
OpenHands	CLI agent · T3 experimental, headless-only	Multi-provider	OpenHands CLI credentials
Gemini CLI	CLI agent · Go-side only	Gemini	Google account or API key
vLLM	Self-hosted inference gateway	Your own models, your GPUs	Endpoint URL; API key optional
llm-d	Distributed inference gateway	Your own models, your cluster	Endpoint URL; API key optional
LiteLLM	OpenAI-compatible gateway	Anything behind your LiteLLM endpoint	Endpoint URL + API key
watsonx.ai	OpenAI-compatible gateway	IBM Granite and configured watsonx models	IBM Cloud API key + project ID
OpenRouter	OpenAI-compatible gateway	OpenRouter model catalog	OpenRouter API key
Custom OpenAI-compatible	OpenAI-compatible gateway	Configured endpoint models	Endpoint URL + optional API key

Active Hives

Want to help? Pick a public hive below and click its Contribute button to run its agents through ClankeR and contribute fixes to a hive of your choice. Your work shows up as tasks completed under your GitHub username, on that hive and on the leaderboard.

Loading hives...

Cross-Hive Leaderboard

Loading leaderboard...

References

Explore the ecosystem

Source code

source repository

ACMM Paper

Autonomy framework

Architecture

How hive works

Hive Commons

Project home

Get started

Request a hive

gh-aw

GitHub Agentic Workflows positioning

Symphony

OpenAI orchestration positioning

vibe-kanban

agent-work coordination UI

Goose / AAIF

Agentic AI Foundation positioning

Reading List

From the Hive Commons Substack.

View all on Substack →

Works with the agents and inference engines you already use

Hive can drive the current T1/T2/T3 agent backends, OpenHands included as new experimental, plus Spektacular (Spek) for long-running run planning. This list describes software integrations and does not imply endorsement or partnership unless marked Partner. TypeSafe AI is a Hive Commons partner.

See all integrations →

Work sources

Agent CLIs by support tier

T1 core: claude, litellm, copilot, codex, goose. T2 supported local path: claude/litellm, codex, copilot, muse, opencode. T3 experimental: goose local, agy, bob, pi, aider, kilo, omp, openhands. Gemini is Go-side only.

A co-maintainer that never sleeps

More than a terminal agent

Trust it a little or a lot. You decide.

Proof, not promises

Start with one. Manage hundreds.

A swarm working on your behalf

Self-host it or onboard in minutes

Plug in wherever your work already lives

The ADLC platform you own outright

How Hive works — and why not build your own

Coverage first, then autonomy

Issue filed

Triage

Fix agent

Coverage gate

Review

Merge

Rerun

Automate the mundane, reclaim your time

Get a hive for your project

Contribute compute to a hive

A pipeline, not a prompt

Running in minutes

Login with GitHub

Install the Hive App

Choose your autonomy level

Connect your AI provider

Agents start working

Conservative by default

Runs the AI you already have

Active Hives

Cross-Hive Leaderboard

Explore the ecosystem

Reading List

Works with the agents and inference engines you already use

Work sources

Agent CLIs by support tier

Planning and runs

Inference engines

Model gateways and providers

Thanks to the organizations that power Hive Commons

Contribute to