Document additional model provider setup paths

[kazimakgul]

Goal

Expand the docs for configuring model providers used through the AgentOS and OpenClaw workflow.

Scope

• Document provider setup expectations
• Add examples for common provider paths
• Clarify where provider readiness is surfaced in AgentOS
• Mention how users can verify setup with agentos doctor

Notes

This is a good first documentation issue.

[airouter-dev]

One provider path worth documenting is AI Router as a hosted OpenAI-compatible endpoint.

• Site: https://ai-router.dev
• API base URL: https://api.ai-router.dev/v1
• Example model: gpt-4.1-mini
• New-user credit: 20U total, with 5U available on signup and 15U unlocked after top-up

Minimal env shape:

OPENAI_API_KEY=your_ai_router_key
OPENAI_BASE_URL=https://api.ai-router.dev/v1
OPENAI_MODEL=gpt-4.1-mini

Minimal smoke-test shape:

from openai import OpenAI

client = OpenAI(
    api_key="${OPENAI_API_KEY}",
    base_url="${OPENAI_BASE_URL}",
)

resp = client.chat.completions.create(
    model="${OPENAI_MODEL}",
    messages=[{"role": "user", "content": "hello"}],
)

For agentos doctor, I would explicitly verify:

• key present
• base URL includes /v1
• model discovery via /v1/models if your doctor flow supports it
• one short chat-completions probe

This keeps the example provider-neutral at the client layer while still giving users a hosted OpenAI-compatible path.