Start free. Pay for more work running at once.
Detent Cloud prices each organization by how many runs it holds at the same time. Every plan, Free included, is the whole product: your runners, unlimited projects, members, repositories, and runners.
One product. The tiers differ in concurrent work.
Concurrent work is the number of runs your organization has going at once. Each square below is one slot.
Free
Try Detent on real work. No card, no Stripe customer.
Every organization starts on Free. Owners upgrade from the organization's billing screen, and checkout runs on Stripe. Downgrading keeps your data and lets in-flight work finish. Prices are USD per organization per month.
Free is not a trial.
Every plan
- Native execution on runners you enroll
- Unlimited projects, members, repositories, and runners
- Up to 10,000 open issues per organization
- The web app, MCP access, and the same gates and serialized landing
Paid plans add
- More concurrent work: 8, 20, or 50 runs at once, or a custom allowance on Scale
- Hosted artifact storage, so run outputs outlive the runner that made them
Paid separately
- Model usage. Runners use the Codex or Claude Code sign-in already on the machine, so agent work bills against your own subscription or API key.
- Luna chat and hosted AI, which draw on AI credits rather than your plan.
- The machines your runners run on.
How concurrent work is counted.
- What counts
- Every running attempt in the organization, whatever its stage. An issue waiting in a lane does not hold a slot.
- At the limit
- The next claim waits. The issue stays in its lane and is picked up as soon as a slot frees; nothing is refused when you file it.
- Open issues
- Every plan allows up to 10,000 open issues per organization. Done, Cancelled, and archived issues do not count.
- Projects, people, runners
- No plan limits projects, members, repositories, or runners. Enroll as many machines as you like; the plan caps how many runs go at once.
Or run the Hub yourself, free.
The Hub is the same binary. Self-hosting needs no Detent Cloud account, subscription, or billing, and no plan limits apply unless you configure them. You run one host and keep it backed up.
From Detent's capacity catalog as of October 22, 2026. Hosted capacity allowances →
The Hub and runners, the state machine, the gate model, worktree isolation, and serialized landing.
The interaction-model argument, the default choices, and the comparison matrix.
Install a runner or your own Hub: platform-tabbed commands, requirements, and the first run.
The operator surface: lanes, agent activity, budgets and capacity, and MCP access.
Reviewed recordings of Detent moving real work through checked-in workflows.
The curated operator reference, mirrored from a pinned Detent release.
FSL-1.1-ALv2, the repository, releases, contributing, and how Detent builds itself.