Raycast

Raycast AI adds UberLLM as a custom provider in providers.yaml. Custom providers need Raycast Pro.

Before you start

• An UberLLM account with credit, and an API key from Dashboard → Keys (ull_...).

Setup

1. Open Raycast Settings → AI → Models & Providers → Custom Providers and click Reveal Providers Config. This opens ~/.config/raycast/ai/.
2. Copy providers.template.yaml to providers.yaml (or open providers.yaml if you already have one).
3. Add UberLLM:

providers:
  - id: uberllm
    name: UberLLM
    base_url: https://api.uberllm.dev/v1
    api_keys:
      default: ull_your-key
    models:
      - id: deepseek/deepseek-v4.1-flash
        name: DeepSeek V4.1 Flash
        context: 1048576
        abilities:
          temperature: { supported: true }
          system_message: { supported: true }
          tools: { supported: true }
      - id: google/gemini-2.5-flash
        name: Gemini 2.5 Flash
        context: 1048576
        abilities:
          temperature: { supported: true }
          vision: { supported: true }
          system_message: { supported: true }
          tools: { supported: true }
      - id: z-ai/glm-5.2
        name: GLM 5.2
        context: 1048576
        abilities:
          temperature: { supported: true }
          system_message: { supported: true }
          tools: { supported: true }

Raycast reads the file again when it changes; no restart.

• base_url stops at /v1. Raycast adds /chat/completions.
• Raycast does not load models from /v1/models: list each one. Any chat id from the catalog works.
• Only mark an ability as supported if the model has it. vision is right for google/gemini-2.5-flash, not for z-ai/glm-5.2.

Recommended models

Checked on 2026-10-30 through UberLLM: the three above.

Common errors