Step 1: Get an API Key
Create a key (sk- prefix) at the API Key console. All requests authenticate via:
Authorization: Bearer YOUR_API_KEYStep 2: First chat request
The platform is fully OpenAI-compatible — point your existing SDK at https://api.maitoken.com/v1:
curl -X POST https://api.maitoken.com/v1/chat/completions \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-5.6-sol", "messages": [{ "role": "user", "content": "Hello, introduce yourself" }] }'Available models come from List Models. Native Anthropic / Gemini protocols are also supported — see the Messages API.
Generate an image (synchronous)
Image generation is synchronous — the response carries the image directly, no polling (keep the client timeout ≥ 180 seconds):
curl -X POST https://api.maitoken.com/v1/images/generations \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-image-1", "prompt": "Cyberpunk city at night after rain", "size": "1024x1024" }'See Image Generation.
Generate a video (asynchronous)
Video generation is asynchronous: submit for a task Id, poll to a terminal state, then fetch the result. Sora example:
# 1. Submitcurl -X POST https://api.maitoken.com/v1/videos \ -H "Authorization: Bearer YOUR_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "sora-2", "prompt": "Sunset over the ocean", "seconds": "8" }' # 2. Poll (every 5-10s until status=completed)curl https://api.maitoken.com/v1/videos/{video_id} \ -H "Authorization: Bearer YOUR_API_KEY" # 3. Downloadcurl https://api.maitoken.com/v1/videos/{video_id}/content \ -H "Authorization: Bearer YOUR_API_KEY" --output video.mp4Each video vendor keeps its native protocol and entry point: Sora / Kling / Seedance.
Next steps
- List Models — all models and protocols available to your key
- Rate Limits — limits and retry guidance
- Tool integrations: Claude Code / Codex