Connecting Raycast AI to an OpenAI-Compatible Gateway

说明与用途

Raycast 用户现在可以直接在启动器中选择 CometAPI 模型，无需打开其他应用，通过在 providers.yaml 中配置自定义提供商并使用 curl 验证连接实现。

问题

Integrating a custom OpenAI‑compatible gateway into Raycast AI so users can select models directly from the desktop launcher without opening another app.

结果

Raycast users can now select and use CometAPI models directly from the launcher, enabling quick AI tasks without leaving the desktop environment.

工作流

• 1. Verify macOS, recent Raycast, and a valid CometAPI API key.
• 2. Test connectivity with curl: curl --fail-with-body --silent --show-error 'https://api.cometapi.com/v1/models' --header "Authorization: Bearer $GATEWAY_API_KEY".
• 3. Open Raycast Preferences → AI → Custom Providers → Reveal Providers Config.
• 4. Copy the template providers.template.yaml to providers.yaml in the revealed directory.
• 5. Add a provider entry: set id, name, and base_url: https://api.cometapi.com/v1.
• 6. Add the API key via Raycast’s secure credential field or environment variable.
• 7. (Optional) Let Raycast discover models via GET /v1/models or list models manually in the config.
• 8. Reload or restart Raycast, then refresh models if available.
• 9. Open Quick AI, select the new provider’s model, and send a short prompt to test.
• 10. Debug issues: check file location, token validity, model discovery, streaming behavior, and latency.
• 11. Optimize workflow: choose lightweight models for quick tasks, higher‑capacity models for deep reasoning, track usage, and create reusable AI Commands.

证据

I’d test authentication and model-list access before touching Raycast. With your token already available locally as GATEWAY_API_KEY, this requests the OpenAI-style models endpoint:

来源

Open Raycast Preferences → AI → Custom Providers → Reveal Providers Config.

来源

Add the token through Raycast’s custom API-key or secure credential fields where supported.

来源