Pricing

Simple plans. Model costs stay on your own key.

Handoff never resells model tokens. Whatever you pay your provider is what authoring costs; replays cost nothing. Plans are priced on runs, browser minutes and seats.

Free

$0 / month

For trying the workflow on one staging app.

  • 50 runs per month
  • 120 browser minutes per month
  • Up to 3 members
  • Live view, takeover, recorded steps
  • Deterministic replay and Playwright export
  • Requirements, plans, cycles, incidents
Create an account

Enterprise

Custom · annual

For regulated teams and larger organisations.

  • Unlimited runs, minutes and members
  • Everything in Team
  • Dedicated or self-hosted deployment
  • Isolated browser tier per organisation
  • Security review and DPA
  • Priority support with a named contact
Contact sales

Limits are enforced per calendar month and shown live under Settings → Usage. A run is one execution of one test case, in agent or replay mode. Browser minutes count from session start to session end.

Pricing questions

What does a typical authored test cost in model usage?
It depends on the model and the page. As a reference, a ten-step flow on a mid-tier model with prompt caching lands around forty cents on the provider's bill; without caching it is roughly three times that. Handoff lays prompts out for cache hits and reports cache-read tokens per call so you can see it working.
Do replays use any model at all?
No. Replay uses recorded locators and assertions only. A test in our own suite asserts that zero model calls happen on that path.
Can we export our tests before deciding?
Yes. Playwright export is available on every plan, including Free, at any time.
Is there an on-premises option?
Enterprise includes dedicated or self-hosted deployment. The stack is Node, Postgres, Redis, Temporal, S3-compatible storage and Chromium, and runs on plain Linux servers without Kubernetes.
Pricing · Handoff