Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
A unified API gateway that routes requests to 70+ AI models from providers including OpenAI, Anthropic, Google, Alibaba, Deepseek, and xAI. Uses a single API key and OpenAI-compatible endpoints, replacing per-provider credentials and SDKs. Supports chat completions, streaming, vision, and function calling.
One API key and one base URL replaces separate credentials, SDKs, and billing accounts for every AI provider.
When to use it
- Switching between GPT-4 and Claude without changing SDK code
- Routing cheap queries to fast models and complex ones to GPT-4
- Analyzing images using GPT-4o through a unified endpoint
- Running A/B tests across multiple models on the same prompt
- Setting up automatic fallback when a primary model fails
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Compare GPT-4.1, Claude-3-Sonnet, and Gemini-2.0-Flash on the same coding question
- 1Set AISA_API_KEY environment variable
- 2Call compare endpoint with models list and the coding prompt
- 3Send identical messages to all three models through api.aisa.one/v1/chat/completions
- 4Collect responses with latency and cost metadata from each
- 5Return side-by-side results with usage.cost per model
Three model responses with latency, token counts, and per-request cost for each
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.