AMD MI300X · 10:00–13:00 PST · daily

EXEA
RESERVE

Book a daily window on the collective node. Point us at a public repo, tell us how to run it in one file, and we run it for you — unattended, three hours a day, resuming from your own checkpoint until it finishes.

Bookings open Aug 8 · reserve up to 60 days out · up to 5 slots each

How it works
1 · Prepare
Make your repository public and add exea.md at its top level — five headings telling us how to set up, start, resume, stop, and save your job. Template →
2 · Connect
On your dashboard, save the repo URL and a fine-grained GitHub token so the runner can clone and push on your behalf. The token is encrypted and never shown again.
3 · Book
Claim up to five days. Multiple people share each day’s window — the runner works the queue across the node.
4 · 10:00 PST
The runner wakes, checks the queue, clones your repo, restores yesterday’s snapshot if there is one, and runs your START or RESUME command.
5 · 13:00 PST
Your STOP command flushes a checkpoint, your SAVE command pushes results, and the disk is snapshotted. Tomorrow it picks up exactly there.
6 · Watch
Progress, session count, and any errors land on your dashboard. Failures come back with the stage and the log tail so you can fix them yourself.
The one thing you must build

Your job is stopped at 13:00 every day. It has to write its progress to disk and pick that progress back up the next morning. If it cannot, it starts from zero every day and never finishes — and work that was never checkpointed is gone. Exea Labs is not liable for lost work. Read the full terms →

Start here

Write your exea.md, then claim a slot.

View slots