Free
For trying ReqRun and running one small app.
- 1 project
- 10,000 requests / month
- 100 pending requests / project
- 2 API keys
- 7-day retention
- BYOK OpenAI key
- Community support
Pricing
Hosted plans include the ReqRun reliability layer. Connect your own OpenAI key per project so OpenAI usage stays billed directly by OpenAI.
Free
For trying ReqRun and running one small app.
Builder
For real usage across multiple apps.
Pro
For production scale and heavier request volume.
Hosted ReqRun subscriptions are billed in advance and handled through Paddle. Paid plans are governed by the Terms of Service, Privacy Policy, and Refund Policy.
No. ReqRun pricing covers request execution reliability. OpenAI usage is billed by OpenAI through the key you connect to each project.
Yes. Local evaluation is useful for development and server testing, but the production commercial product is hosted ReqRun with managed operation, limits, and support.
No. The launch target is developer-friendly self-serve usage with clear limits.
A request run is one accepted ReqRun request, whether it completes immediately or finishes later through the worker.
ReqRun warns at 80% usage and pauses new work at the monthly limit. It also caps pending work per project so one app cannot flood the queue. Builder removes the Free blocks with more projects, more API keys, more monthly requests, and a larger pending queue.
The SDK, examples, and local evaluation path can be developer-friendly without making the full hosted control plane the free production product. Licensing and self-host terms will be explicit before any production self-host offer.
ReqRun does not offer prorated refunds for partial periods or partial quota usage once request runs have been sent. Misuse suspensions are not refundable. See the Refund Policy for the full rule set.
Possibly. These are early-access hosted plans and limits while the product is still launching.