Documentation

Everything you need to ship software with a fleet of autonomous agents — from your first product to a landed stack of reviewed PRs.

Quickstart

  1. Sign up (free, no card) — you land in the Agents home.
  2. Create a Product — an outcome you want built.
  3. Connect a repo: a GitHub repo (via the GitHub App) or a local path.
  4. Type a goal in the product's Ship-it box and hit Ship it.
  5. A fleet of agents fans out, builds, and opens PRs you review and land.

Self-hosting? Jump to Self-host.

Products & repos

A Product is an outcome (not a board). Each product has one or more repos — a GitHub repo you keep on GitHub (we drive it: clone, branch, commit, PR, merge) or a local path for desktop / offline work. Agents run against the product's primary repo.

  • GitHub — connect via the GitHub App; work lands as pull requests.
  • Local — point at a path; work runs in isolated git worktrees.

Ship a goal

Describe an outcome. A planner decomposes it into missions, and the bounded dispatcher fans them out to Builder agents — in parallel, on the BEAM — each in its own worktree. Missions flow through four stages as their runs progress.

Add Apple Pay to checkout, behind a feature flag…
Ship it →
Stack the missions — each PR builds on the previous, reviewed in order
Queued
Refactor cart service
Working
Add Apple Pay · ⚡
Review
Cart totals API · PR #128
Landed
Payment schema · merged

The product flow view — missions grouped by derived stage.

Stacked PRs

Tick “Stack the missions” in the Ship-it box and a goal becomes a graph of dependent PRs. The planner says which missions actually build on which; every mission with no upstream starts at once, and each downstream one is released as its upstreams push their branches. They land upstream-first through the merge queue, and GitHub auto-retargets the rest on merge.

  • A dependency graph, not a queue — two independent missions run in parallel, and a mission that needs both waits for both.
  • Cards show ⤷ stacked on <branch> so the stack is legible.
  • Small, reviewable PRs instead of one giant diff.
Payment schema
af/payment-schema
Cart totals API
af/cart-totals
⤷ stacked on af/payment-schema
Add Apple Pay button
af/apple-pay
⤷ stacked on af/cart-totals

Review & land

Agents don't merge to production on their own. Completed runs land in the Review Inbox; you review the diff, then the merge queue verifies each one behind a gate and lands it serially (a PR on GitHub, a branch merge locally). Landing is dependency-ordered for stacks.

Skills

Skills are reusable SKILL.md playbooks — the knowledge an agent draws on for a kind of work. Browse the library at /skills, enable the ones your org's agents may use, or import any MIT repo from GitHub in a click (admin-only; playbook text only, bundled scripts flagged, never run).

Models & cost

Every seat runs the 80/20 router: routine work runs on local / open models (free, private, via Ollama), and only the hard ~20% escalates to a frontier model like Claude Opus — metered against your per-org budget. Bring your own key, or your own Ollama. See /models.

Integrations

Agents act through a governed MCP boundary — every tool call is authorized against your policy before it runs. Browse 100+ MIT-licensed MCP servers at /integrations, and bring your own MIT server too.

Governance & memory

Governance: each agent credential has a role; role-based policy controls which tools and skills it may invoke, who can trigger a run or a merge, and per-org spend caps — beyond what branch protection covers.

Memory: agents remember decisions and conventions across runs (org, product, and mission scope) — injected into every prompt and captured after each success, so the fleet improves over time.

Self-host

Agentic Flow is MIT-licensed and portable. Clone it, run it yourself, keep your data and models in your environment.

git clone https://github.com/vinpatel/agenticflow
mix setup
mix phx.server   # → http://localhost:4000
Start free →