Work on Repeat
Menu

Simple monthly plans

Start one routine free. Put the rest into production for $39.

Every plan includes model choice, explicit tool controls, and inspectable run history. Upgrade when recurring work becomes part of your operation.

Free

$0

No card required

Prove one recurring responsibility before you pay.

  • 1 active routine
  • 10 run starts each UTC month
  • 1 MCP connection
  • 7 days of run history
  • 1 run at a time
  • 3 model steps per run
  • Read-only developer API
  • $0.50 of AI usage each month
Put one routine to work

Recommended

Pro

$39

per month

Run several responsibilities with API access and signed delivery.

  • 10 active routines
  • 250 run starts each billing month
  • 10 MCP connections
  • 90 days of run history
  • 2 concurrent runs
  • 12 model steps per run
  • Full developer API
  • Signed result webhooks
  • $10 of AI usage each billing month
Put recurring work into production

Business

Contact sales

Scoped to your operation

For higher volume, longer history, and agreed operating requirements.

  • Routine, run, and connection limits configured for your workload
  • Longer run-history options
  • Full developer API and signed result webhooks
  • AI usage configured per agreement
  • An agreed support path
Talk through your workload

Monthly billing. When paid access ends, Free limits apply, including 7 days of accessible history.

Run starts

Count work, not clicks.

Manual, API, and scheduled starts share one allowance. Invalid configuration never consumes a run, while an execution that reaches the model does.

AI usage

Model cost in dollars.

Usage reflects reported upstream model cost. Free includes $0.50 each UTC month; Pro includes $10 each billing month.

Cost control

A limit, not an overage.

We pause new starts when an allowance is exhausted. Nothing is deleted, and there is no surprise usage bill at launch.

Before you choose

Clear limits. Predictable next steps.

Free is enough to prove one real responsibility. Pro gives an operating workload room to repeat without turning pricing into a credit puzzle.

What counts as a run start?

A run start is reserved when a manual, API, or scheduled run is accepted for execution. Validation errors before enqueue do not count.

What happens when a run fails?

Once model execution begins, a failed run still uses one run start because upstream work may already have happened. Any reported model cost also counts toward AI usage.

What happens when I reach a limit?

New run starts pause at the run or AI-usage allowance. Your routines, settings, and prior results stay intact. Upgrade to Pro or talk with us about a Business workload; there are no automatic overage charges.

What happens when I cancel?

Pro stays active through the paid period. After that, the workspace returns to Free limits. Records stay intact, but results are accessible only within Free's 7-day accessible-history window. Schedules above the Free allowance pause until you choose what to resume.

How does AI usage work?

AI usage is the upstream model cost reported through LLMGateway. It is measured in dollars against the allowance for the plan's usage period—no invented credits and no surprise overages.

Is tax included?

Published prices exclude tax. Stripe calculates applicable tax from the billing details supplied during Checkout.

When should I talk to you about Business?

Business is the right conversation when you need more than 250 monthly run starts, longer than 90 days of accessible history, or operating and support requirements that need to be agreed explicitly.