Simple, per-organization pricing
Every plan includes the full product — live runs, workflows, personas, all trackers and notifiers. Tiers only change how much you can run.
- 25 runs per month
- AI tokens included during the beta
- 1 concurrent run
- Up to 10 members
- All trackers & notifiers
- Gated workflows & live view
- Community support
- 250 runs per month
- 500M AI tokens per month
- 3 concurrent runs
- Up to 25 members
- All trackers & notifiers
- Published artifacts on every surface
- Email support
- 1,000 runs per month
- 2B AI tokens per month
- 10 concurrent runs
- Up to 100 members
- All trackers & notifiers
- Remote runners with labels
- Priority support
- Custom run volume & concurrency
- Self-hosted deployment support
- Your runners, your network
- Security review & SSO (roadmap)
- Dedicated support channel
Peractor is in beta — paid tiers are provisioned manually while billing is finished, so "talk to us" really means talk to us.
Questions
What counts as a run?
One execution of a workflow against a task — however many phases it has. Reworks after "request changes" continue the same run and don't count again.
What are AI tokens?
The model usage Peractor pays for: runs executed on our infrastructure, the in-app assistant, and prompt improvement. Every plan includes a monthly allowance, shared across all three and counted per organization. You can watch it in Settings.
Can I bring my own agent credentials?
Yes — by bringing your own machine. Pair a runner in Settings and its runs execute on your hardware with your Claude (or Codex) login, billed by your model provider at cost and counted against no allowance of ours. The assistant and prompt improvement always run on Peractor's own credential.
Is there a per-seat charge?
No. Members are unlimited on every tier — reviewing your agents' work should never cost extra. Pricing is per organization.
Can I self-host Peractor?
Peractor is MIT-licensed open source, and the hosted product is the same codebase. If you want to run it yourself, the repository ships Docker Compose files — Enterprise support covers the rest.
What happens when I hit a limit?
New runs are declined with a clear message until the month rolls over (UTC) or you upgrade — the same for runs, for tokens, and for the assistant. Nothing in-flight is interrupted, and your data is never held hostage.