The autonomous improvement platform

Verified autonomy
for your codebase.

An agent fleet runs a verified build → verify → fix loop on your repos and ships merge-ready PRs on your own provider keys. Unlike feature-autopilot tools that hand you unverified diffs, automaintainer only ships work that passed its gate.

Free forever · No credit card · You stay in control
am — agent run
$ am repo add https://github.com/acme/backend \
--focus "fix failing tests, bump outdated deps"
✓ Added repo backend (a3f291)
$ am repo enable a3f291
✓ Enabled. Next run: tonight 2:00 AM
— 6 hours later —
↓ architect · analysing 47 files...
↓ dev · fixed 3 tests · bumped 8 deps
↓ qa · LGTM · opening PR #142
✓ PR merged · 2 issues closed
Staff augmentation

Point it at the work
you keep putting off.

Write a focus in plain English — or pick a playbook. The agent team scopes it, does it, verifies it passes, and opens a reviewable PR — every run, on your schedule.

Keep tests green Raise coverage to 80% Bump dependencies safely Fix lint & type errors Harden error handling Keep docs in sync Close stale GitHub issues

If any of this sounds familiar…

“The debt pile grows faster than we can pay it down.”

Your 2–3 senior devs are constantly context-switching between features, infra fires, and tech debt. Tests are a prayer. Docs are lies. Nobody has sprint capacity to fix it.

“I can’t hire fast enough — and contractors take 2 weeks to ramp up.”

Junior devs need oversight that costs more time than it saves. Senior contractors are expensive and still need context. You need output, not onboarding.

“PRs are piling up. Review bandwidth is the real bottleneck.”

You have open PRs from last sprint that nobody’s looked at. Every hour unreviewed is an hour of merge-conflict debt accumulating silently.

AM was built for exactly this. Here’s what it can take off your plate right now — and where it’s headed next.

Now — the maintenance & quality quadrant

These are the tasks AM is already doing reliably in production — bounded, verifiable, repeatable work that your team keeps deferring.

  • Tests
    Add missing tests, fix failing ones, raise coverage to a target — verifiably.
  • Dependencies
    Bump outdated packages, validate nothing broke through the verify gate.
  • Lint & type errors
    Systematic cleanup runs — not one-off, on a cadence.
  • Docs
    Keep README, API docs, and inline comments in sync with the actual code.
  • Small refactors & issue triage
    Bounded, scope-controlled refactors. GitHub issues fixed and closed automatically.
The maintenance/quality quadrant is where AM earns trust — and where your team’s time is currently disappearing.
Next — the context & trust quadrant

The feature/architecture quadrant requires context depth AM doesn’t have yet. Here’s what’s on the roadmap to get there:

  • Roadmap-aware agents
    No-touch zones, deprecation markers — AM stops touching code you’re planning to replace.
  • Intent-aware PR review
    Beyond “does it build” — does this PR fit the architecture and the team’s direction?
  • Objective decomposer
    “Reduce checkout latency” → scoped, verifiable sub-tasks AM can actually execute.
  • Architecture context ingestion
    Feed ADRs, design docs, style guides — so AM reasons about your conventions, not generic ones.
The gap is context depth and trust-building. We’re closing it — starting from a strong verified foundation in the maintenance quadrant.
Everything you need

A fleet that verifies
before it ships.

A fleet of agents — verified, not vibes

Specialised agents — architect, dev, and QA — coordinate over a peer message bus, then every change runs the Verification Gate (build → test → fix → re-verify). Only work that passed its own gate becomes a PR.

architect
analysing codebase...
dev
3 commits ready
qa
APPROVE ✓

Scheduled or on-demand

Set a cron-style window. Or trigger a run instantly from the CLI or REST API.

days: [Mon–Fri]
hours: 02:00–06:00
next run: 3h 22m

Review & auto-merge

A separate review team reads diffs and checks CI on every PR (Pro). On Max it merges approved PRs for you — fully hands-off.

GitHub issue sync

Open issues are injected into every run. Agents fix them and close them automatically with Fixes #N commits.

Agent-first CLI & API

JSON by default. Semantic exit codes. Structured errors. Designed for AI agents to drive, not just humans.

0
manual steps after setup

In hands-off mode (Max): from repo registration to merged PR with no human in the loop. On Free & Pro you review before merge.

Zero collision

Git worktree isolation

Each run gets a dedicated branch and isolated worktree. Two concurrent runs on the same repo can never overwrite each other — even if the previous run never finished.

~/.automaintainer/repos/a3f291 # shared base clone
~/.automaintainer/work/am-a3f291-x9k2p ← run A
~/.automaintainer/work/am-a3f291-y7m4q ← run B (safe)
branches: am/am-a3f291-x9k2p · am/am-a3f291-y7m4q
Simple by design

Up in 3 steps.

1
Sign up and add your repo via web dashboard
Sign up → Add repo URL
Set focus: "keep tests passing, bump deps"
2
Install the worker & plug in your own AI key (BYOK)
am-cloud start --panel wss://... --token <key>
# harness uses YOUR provider key — you pay tokens directly, no markup
3
Set schedule and walk away
Configure schedule in dashboard
# Worker runs automatically. PRs appear. Issues close.
Five building blocks + memory

Loop engineering,
already wired.

“You shouldn’t be prompting coding agents anymore. You should be designing the loops that prompt them.” The catch: today that means wiring five primitives together yourself and babysitting the glue. Automaintainer ships them assembled — one hosted, verified loop across a fleet of workers and models.

1 · Automations
The heartbeat

Scheduled runs + explore-mode discovery find the work and triage it on a cadence. Nobody is pressing go.

2 · Worktrees
Parallel without collisions

Every run gets its own isolated git worktree, so agents working in parallel never step on each other’s files.

3 · Skills
Intent, written down once

Playbooks + per-repo conventions capture project knowledge, so the loop compounds instead of re-guessing from zero each run.

4 · Connectors
Acts in your real tools

Opens PRs, closes issues, runs your verify command, and (via MCP-capable harnesses) reaches the rest of your stack.

5 · Sub-agents
Maker ≠ checker

Architect / dev / QA teams — and a separate verifier from the maker, so no agent grades its own homework.

+ Memory
The agent forgets, the repo doesn’t

Per-repo memory + a run ledger live on disk, outside any conversation — so tomorrow’s run picks up where today’s stopped.

Build the loop. Stay the engineer.

A loop running unattended is a loop making mistakes unattended — unless “done” means something. AM’s verification gate is a different agent from the maker, and a red gate never ships. You decide review vs. auto-merge. No slop, no cognitive surrender.

BYOK · harness-agnostic

One loop.
Many minds.

Every model is trained differently — so put the right one on each seat. A frontier model to architect, a strong coder to implement, a fast skeptic to verify. Mix tiers to balance quality against token cost, and swap any of them anytime. Your keys, no markup.

Fable 5 Opus Sonnet DeepSeek Mimo + any model you bring

Just shipped — Mythos’ Fable 5: assign it to any role the day you have a key.

Pricing

Starts free.
Scales with you.

You pay for orchestration — not inference. automaintainer is bring-your-own-keys: your worker runs the AI harness with your own provider key, so you pay tokens directly with no markup. That's why it's €19, not $200+.

Free
€0/mo
You review & merge
  • 1 repo
  • 3 runs / day
  • 1 worker
  • GitHub issue sync
  • AM code review
Get started free
Pro
€19/mo
AM reviews; you merge
  • 5 repos
  • 30 runs / day
  • 2 workers
  • GitHub issue sync
  • AM reviews your PRs
Start Pro trial
Max
Popular
€49/mo
Hands-off after the gate
  • 25 repos
  • 150 runs / day
  • 5 workers
  • AM review + auto-merge
  • GH issue → PR auto-close
  • Priority support
Start Max trial
Enterprise
Custom
Volume & SLA
  • Unlimited repos, runs & workers
  • Everything in Max
  • Self-hosted deployment
  • Custom AI model / CLI
  • SSO & audit logs
  • Dedicated SLA
Contact sales

It doesn’t just write code.
It verifies it, then ships it.

Add your first repo in under 60 seconds. Maintenance wedge today — autonomous improvement on your schedule.

Get started free — no card required

Need a human in the loop? The founder is available for architecture reviews, custom integrations, and team ramp-up. Check pricing at Intrane →