Report issues and request integrations

Use the public AnyRouter issues repository to report bugs, request features, and propose new upstream provider integrations.

Use the public AnyRouterHQ/issues repository to report problems, request features, and propose new upstream provider integrations. The repository includes issue templates that help you share the details we need.

Choose the right issue type

If you want to…Open this templateInclude
Report a bugBug reportWhat happened, steps to reproduce, expected and actual behavior, affected endpoint or page, model ID, request ID or timestamp, and sanitized logs
Request a featureFeature requestThe problem or use case, your proposed solution, alternatives, impact, and examples
Request an upstream/provider integrationUpstream provider integrationProvider details, API compatibility, model catalog information, and the models you want integrated

Before you open an issue

  • Search existing issues to avoid duplicates.
  • Check status.anyrouter.dev for active incidents before reporting an outage or error spike.
  • Remove secrets and personal data. Never post API keys, provider keys, tokens, passwords, or private customer data.
  • For account-specific, billing, or security-sensitive matters, email support@anyrouter.dev instead of opening a public issue.

Report a bug

Use the bug report template for problems with the API, dashboard, model routing, docs, CLI, MCP integrations, or website.

Include details that help us reproduce the issue:

  • The affected area, endpoint, or page URL.
  • The model ID, such as openai/gpt-5.4-mini.
  • The request ID, approximate timestamp, and timezone for API requests.
  • A minimal request example or reproduction steps.
  • The full error response or status code, with secrets removed.

Request a feature

Use the feature request template when you want AnyRouter to add or improve a capability. Describe the problem you are solving, who it affects, and what a good solution would look like. Links to examples, API shapes, or similar implementations help us evaluate the request.

Request an upstream provider integration

Use the upstream provider integration template when you want AnyRouter to add a new provider or provider API. Complete, verifiable information helps us review compatibility and model coverage.

Provide the following information:

FieldWhy it is neededExample
Provider nameIdentifies the provider to integrateExample Provider
Provider websiteConfirms the official sourcehttps://provider.example
Provider descriptionExplains what the provider offers and why it should be integratedAPI for coding and reasoning models
Logo URL or SVGRequired for provider and model display in AnyRouterhttps://provider.example/logo.svg
API base URLShows where requests are senthttps://api.provider.example/v1
Model listing/catalog pageRequired to verify available models and IDshttps://provider.example/models
Requested modelsLists the exact models you want, matched to the catalogprovider/model-a, provider/model-b

API compatibility details

Select and describe each supported API surface:

  • OpenAI-compatible Chat Completions API (/chat/completions or equivalent): include the endpoint path and docs link.
  • Anthropic-compatible Messages API (/messages or equivalent, optional): include the endpoint path and docs link.
  • OpenAI-compatible Responses API (/responses or equivalent, optional): include the endpoint path and docs link.
  • Embeddings API (optional): include the endpoint path and docs link.
  • Decision / System One API (optional): include the endpoint path and docs link.
  • Other APIs: describe the protocol, endpoints, and documentation.

Also include authentication details (header or key format, without secrets), pricing or limit notes, terms that affect integration, and any sample requests or special behavior such as streaming headers.

After you submit

A clear issue helps maintainers confirm the problem, reproduce it, and evaluate the request. If you have more information later, add it as a comment instead of opening a duplicate issue.

AnyRouter routes requests to third-party models. Models and outputs are provided as is, with no warranty; you use them at your own risk. Disclaimer