Frequently Asked Questions

Visit our registration page and create an account. After signing up, navigate to the dashboard to generate your API key. New users can use promo code DEVSTARTER20 to receive ¥20 free credit.

Yes! Our API is fully compatible with the official OpenAI SDK for Python, Node.js, and other languages. You only need to change two things: set base_url to our endpoint and use your AI Token API key. All existing code using the OpenAI SDK will work without modification.

Migration is simple — just update your client configuration:

# Before (OpenAI)
client = OpenAI(api_key="sk-...")

# After (AI Token)
client = OpenAI(
    api_key="your-aitoken-key",
    base_url="https://bb71d3f3506709ed-101-126-19-34.serveousercontent.com/v1"
)

That's it! No other changes needed. See our code examples for more details.

Any language that can make HTTP requests. We provide ready-made examples for Python, Node.js, Java, Go, and cURL. Since we follow the OpenAI API spec, all official OpenAI SDKs work out of the box.

We offer 38+ models including DeepSeek V3/V4, Qwen (all versions), GLM-5.2, Kimi K2.7, MiniMax M2.5, and more. This includes text generation, reasoning, code, vision, audio, and embedding models. See our complete model list for details.

Here's a quick guide:

• General tasks: DeepSeek V4 Pro or Qwen 3.5 397B
• Code generation: Qwen 3 Coder Plus or Kimi K2.7 Code
• Budget projects: DeepSeek V3 or Qwen Turbo
• Reasoning/Math: DeepSeek R1
• Image understanding: Qwen VL Max
• Embeddings/RAG: Qwen3 Embedding 8B
• Long documents: Qwen Long (10M context)

DeepSeek V4 Pro is the latest generation with significantly improved reasoning, math, and coding capabilities. V3 is the previous generation — still excellent but at a lower price point. For most applications, V3 provides great value; use V4 when you need the absolute best quality.

You pay per token used. Each model has its own pricing (see pricing page). There are no monthly fees — just buy credits and use them at your own pace. Credits never expire on paid plans. New users get ¥20 free with code DEVSTARTER20.

We accept USDT (TRC20/ERC20), Alipay, and WeChat Pay. Enterprise customers can also use invoice billing. Contact us for payment arrangements.

Yes! New users who register with code DEVSTARTER20 receive ¥20 (approximately $2.80) in free credit. This is enough for thousands of API calls with our budget models.

Free tier credits have no expiration. Paid plan credits also do not expire. Use them whenever you need.

Yes. Set "stream": true in your request body to receive responses as Server-Sent Events (SSE). Each chunk contains a partial response. See our streaming examples.

Rate limits depend on your plan:

• Free: 20 requests/min, 40K tokens/min
• Pro: 60 requests/min, 200K tokens/min
• Business: 200 requests/min, 1M tokens/min

Enterprise plans have custom limits. Contact us for details.

The API returns standard HTTP status codes with descriptive error messages. Common codes:

• 401 — Invalid API key
• 429 — Rate limited (wait and retry)
• 502/503 — Model unavailable (try another model or retry)

We recommend implementing exponential backoff for retries. See the error codes reference.

Our infrastructure includes automatic failover at the provider level. If a specific upstream provider is temporarily unavailable, requests may be automatically routed through an alternative path. For application-level failover, we recommend implementing fallback logic in your code to try alternative models.

Yes! Most flagship models (DeepSeek V4 Pro, Qwen Max, etc.) support function calling via the standard OpenAI tools parameter. See our function calling example.

We offer multiple support channels:

• Email: silkflow@api.service
• Telegram: @aitoken_api
• GitHub: aitoken-hub

Pro and Business plan users receive priority support.

We maintain 99.5%+ uptime. Our multi-provider architecture ensures redundancy. Check our status page for real-time service health.

All API communication is encrypted via HTTPS/TLS. We do not store conversation content beyond what's necessary for processing. API keys are hashed and salted. We follow industry best practices for data security.