Autopilot Cruise

In one line

One goal, broken into steps, run in order — and it survives being killed.

Everything else in Cockpit is the manual toolkit: you pick a mode and drive. Cruise flies the whole task instead. Type a goal in the main prompt box, press ✈ PLAN in the Hive section of the left sidebar, and the result lands back in your chat. You clear the one critical moment; the busywork runs itself.

Enter never changes meaning. It does what it has always done — dispatch to the agent you picked. Planning happens only when you ask for it. There is no mode to remember and nothing hidden that changes what your keyboard does.

What actually happens

A small local model writes the plan. qwen2.5:7b by default, on your own machine — free, and nothing leaves the box. It breaks the goal into 2–5 ordered steps, each with one concrete instruction and a tag for the kind of thinking it needs (text, code, reasoning, web, vision). If that plan comes back malformed, Cockpit automatically re-plans with a larger model (qwen2.5:14b) — so you pay the slower model's time only when the fast one actually failed.
The plan is shown before the work happens. ✈ PLAN prints the steps and runs them straight through. If you would rather read them first, open Chains in the Hive section and press ⛓ Draft a plan from the prompt box: that creates the chain as a Proposal which sits there until you press Run. Use it when the goal is expensive or destructive.
Each step is handed the previous step's output. Step 2's prompt literally begins with a %PREV% marker replaced by whatever step 1 produced. That is the chain.
Every step is a file on disk with its own status — pending → running → done — and its output saved beside it.
The final answer returns to your chat.

Why it survives being killed

Because each step is a file with a status, closing the laptop, crashing, or hitting Ctrl-C does not throw the run away. Finished steps stay done with their outputs intact; the interrupted step flips back to pending; Resume re-runs only that one step. You never redo completed work — which is exactly what most one-shot agent tools lose.

It degrades instead of dying, too: if the planner model is missing or returns nonsense, Cruise falls back to a single step containing your goal verbatim, behaving like an ordinary prompt.

When it pays off

The pattern worth noticing: Cruise pays off most when one step is slow or expensive and the steps after it are fiddly.

Write and file a changelog entry. → collect what changed · draft the entry · insert it in the right place. Kill it during the draft and Resume re-drafts only — the research survives.
Subtitle a video. → transcribe · clean up and segment · translate. The 40-minute transcription is the expensive step; a crash during translation does not re-transcribe it.
Digest your own notes. → extract key points · outline · write the page.

One honest caveat

Small local models sometimes ignore instructions even with the text right in front of them. The planner defaults to a 7b model, which is exactly why the plan arrives as something you approve rather than something that simply runs. In testing it once planned “edit the file” → “check the live site” with no deploy step in between — structurally perfect, and it would have verified a page that never changed. Point it at a larger planner and the plans get noticeably better.

Cruise and AutoCognition

Cruise builds the chain. AutoCognition decides who runs each step, can run some of them in parallel, and can put a step to a vote.

That is also where the tier line falls: planning plus sequential execution is Cruise (Basic); routing, quorum, parallel groups and the message bus are AutoCognition (Pro). Both are open free during the beta.