Autopilot

Autopilot

Have a conversation. Build a plan. Stage the work. Autopilot picks it up, implements it, your CI/CD pipeline fires, and the change is live on a deployed URL. You made the call that mattered — without writing the code.

Autopilot — queue2 runners active
#214

Exponential backoff for Stripe webhook retries

runner mini-pc-01 · worktree wt-3f2a

implementing
#209

Empty state for the deployments page

pushed · [devspec:209] → staging

deployed
#216

Rate-limit headers on the public API

staged behind #214 — shares lib/api/middleware.ts

staged

— staged work, mid-afternoon. Two runners working, one item waiting its turn.

01The Problem

What used to take a sprint takes a conversation

The gap between thinking and shipping has collapsed — but most teams still queue decided work behind keyboards. Someone has to find time, rebuild context, and type out a change everyone already agreed on. That's the bottleneck Autopilot removes.

You're in a session at 2pm discussing a bug. By 2:20 the fix is deployed. You didn't write the code — but you made the call that mattered.

02The Mechanics

How it works

Autopilot runs on your machines, with your existing coding agents — DevSpec orchestrates the staged work, the context, and the record.

Stage decided work

Any open action item can be staged for autonomous execution. DevSpec analyses dependencies and potential conflicts to order the staged work intelligently — and you can rush an item to the top whenever judgment says so.

Fig. i — Staged workapp.devspec.ai/autopilot

Staging knows which items share files — and orders around it.

Runners on your machines pick it up

The DevSpec plugin turns Claude Code on any machine into a runner. Each runner claims an item atomically — no duplicate work — and implements it in an isolated git worktree, so parallel runs never collide.

Fig. ii — Active runnersapp.devspec.ai/autopilot

A spare machine becomes a teammate.

Launch a swarm that never steps on itself

Agent swarms are the hype; collisions are the reality. Flip on Parallel Mode, click an item, and DevSpec instantly shows which other work is safe to run alongside it — it predicts the files each item will touch and keeps a live map of overlaps, shared features, and dependencies. Select a compatible set and launch it in one action: staged for autopilot, or one supervised session per item on your own machine. Anything DevSpec can’t verify is marked unknown — never assumed safe.

Fig. iii — Parallel Modeapp.devspec.ai/agents

Pick one item; everything that can safely run beside it lights up.

Every commit traces back to the conversation

Runners commit with a [devspec:id] tag, and deployment webhooks link the resulting deploy back to the item — and through it, to the session where the team made the decision. Full chain of custody, automatically.

Fig. iv — Chain of custodyapp.devspec.ai/autopilot

Conversation, commit, deploy — one unbroken line.

Failures come back as work, not surprises

When a push breaks the build, DevSpec detects it, reads the failure, and stages a fix item for Autopilot to take another pass. Broken builds become staged fixes instead of morning emergencies.

Fig. v — Build failuresapp.devspec.ai/autopilot

The 3am build failure is staged for a fix by 3:01.

03An Index of Capabilities

Autonomy with a paper trail

01

Worktree isolation

Every run happens in its own git worktree. Parallel runners never trample each other — or your working tree.

02

Atomic claiming

An item is claimed by exactly one runner, transactionally. No duplicate implementations, no racing agents.

03

Live run visibility

Watch each run step by step: which machine, what state, what output. Nothing happens in a black box.

04

Metrics that build trust

Success rates, durations, streaks, and per-type breakdowns. Over time you learn exactly what Autopilot handles reliably — from data, not vibes.

05

Staging intelligence

Dependencies and conflicting changes are factored into execution order, so multiple runners avoid stepping on the same files.

06

Safe swarms by construction

Parallel Mode reads a live conflict map — predicted files, shared features, dependencies — so a swarm is assembled from work that provably cannot collide. No file data means "unknown", never "safe".

07

Your subscriptions, working

Runners use the coding agents your team already pays for — Claude Code today, with the MCP-based protocol open to other runners. Idle capacity becomes shipped work.

Stage the week's small fixes on Friday, leave a runner going, and start Monday by reviewing a stack of pushed branches instead of writing them.

DevSpec

Convinced?

Set up your first project in under five minutes. Connect your repo, start a session, and see what your duck can do.

Start Your Project

14-day free trial · 1,500 credits included