Early Tester Program

llmfaucet is opening in small groups while we test routing, coding-agent compatibility, fair-use limits, and provider reliability. We select testers to cover different tools, operating systems, regions, and real workflows.

What testers receive

• Controlled preview access at https://api.llmfaucet-preview.dev/v1
• A personal, time-limited preview key
• Setup guidance for the preferred coding tool
• Direct feedback channels through GitHub Discussions and Issues
• A chance to influence compatibility and routing priorities

Sponsorship is not required and does not guarantee selection.

What we ask

• Test one real, non-production-critical workflow.
• Report reproducible issues when possible.
• Include client/version, OS, selector, UTC timestamp, request ID, routed-via header, sanitized error, and reproduction steps.
• Respect provider terms, fair-use limits, and the preview access period.
• Never include secrets, credentials, private repositories, or sensitive source code in a report.

Application process

The waitlist application requires GitHub sign-in to reduce spam and bind the application to one account. We ask about your tool, use case, expected volume, operating systems, and ability to report bugs. Applications are reviewed manually; acceptance is not guaranteed.

Preview limits and expiry

Approved early testers receive 100 requests/day, two concurrent streams, and a preview entitlement for 14 days. Capacity, models, routes, and limits can change without notice. At expiry, the account downgrades to the registered 50-requests/day tier instead of breaking the account.

Safe feedback example

Client: Aider 0.x
OS: Linux
Selector: auto:coding
Time: 2026-08-24T12:00:00Z
Request ID: req_redacted
X-Routed-Via: provider/model
Result: 429 after two bounded attempts
Reproduction: one sanitized sentence describing the request shape

Do not paste the prompt, completion, API key, or private repository contents.