feat: add Requesty as an OpenAI-compatible gateway provider

[Thibaultjaigu]

Summary

Adds Requesty, an OpenAI-compatible LLM gateway, as a provider next to OpenRouter. It goes through the existing OpenAI client, so there is no new code path.

Changes

• catchme/services/providers.py: new requesty entry in the gateways block (default URL https://router.requesty.ai/v1, key page https://app.requesty.ai/api-keys)
• catchme/run.py: default model for catchme init (google/gemini-3-flash-preview, same as OpenRouter) and listed under Gateways in the provider picker
• catchme/tests/test_llm.py: default URL test, mirroring the OpenRouter one
• README.md and the es / ja / zh READMEs: row in the supported providers table
• CATCHME-full.md: row in the recommended defaults table

How to test

{
    "llm": {
        "provider": "requesty",
        "api_key": "rqsty-...",
        "model": "openai/gpt-4o-mini"
    }
}

Or run catchme init and pick Requesty under Gateways. Keys are at https://app.requesty.ai/api-keys. EU users can set "api_url": "https://router.eu.requesty.ai/v1".

Validation

• ruff check catchme/ and ruff format --check catchme/: pass
• pytest: 155 passed
• Live check through catchme.services.llm.LLM with provider: requesty and no api_url (so the default URL comes from providers.py): complete() and stream() both returned replies for openai/gpt-4o-mini and google/gemini-3-flash-preview

Disclosure: I work at Requesty. Happy to adjust anything to match project conventions.