FAQ
Answers common usage questions and explains protocols and default models.
Q1: What is this service?
This is an API gateway that provides access to 30+ AI models through standard-compatible protocols. It supports OpenAI-compatible chat completions, Anthropic-compatible messages, and can connect to Claude Code, Cursor, Cherry Studio, VS Code extensions, and other platforms.
Q2: What protocols are supported?
Both OpenAI-compatible and Anthropic-compatible protocols are supported. For Claude-specific tools (Claude Code, Codex), select Anthropic-compatible mode. For standard chat clients, use OpenAI-compatible mode.
Q3: Which models are supported?
The service supports many model families including OpenAI (GPT), Anthropic (Claude), DeepSeek, Qwen, Kimi, Zhipu (GLM), Grok, MiniMax, StepFun, Doubao, Tencent Hunyuan, Meituan LongCat, and Xiaomi MiMo. For current model names, multipliers, and capabilities, see the Model Directory and Pricing Guide.
Q4: How is pricing calculated?
Billing is based on each model's multiplier. The base rate is 1× = $0.055 per 1M tokens (¥0.4 at 1×). Higher-multiplier models cost proportionally more per token. Cache hits are billed at 0.1×, significantly reducing costs for repeated context. See the Pricing Guide for details.
Q5: Do credits expire?
No. Credits never expire or reset. Unused credits remain available indefinitely. You can top up more credits anytime from the dashboard.
Q6: Is caching supported?
Yes. Cache hits reduce overall usage costs significantly. When the same context is reused across requests, cached portions are billed at 0.1× instead of the full multiplier. This is especially beneficial for long conversations, agent workflows, and code completion.
Q7: Can I use the service globally?
Yes. There are no regional restrictions. The service is accessible from any country with an internet connection.
Q8: Are custom plans available for high-volume usage?
Yes. Contact support for custom pricing if you need large volumes or enterprise-level access.
Q9: How do I get my API key?
Register an account, top up credits, then create an API key from your dashboard. Keys are active immediately.
Q10: What if I get 401 / 403 errors?
401 means your API key is invalid or missing. Check the Authorization header (use Bearer format). 403 means your key doesn't support the requested model — check your key type.
Q11: What is claude-sonnet-4-6?
This is our aggregated model optimized for high-concurrency workflows, Python scripts, and text processing. It uses multi-model load balancing for better performance and availability.
Q12: How do I check my usage?
Visit the Balance & Usage page and enter your API key to see real-time usage, token consumption, and cost breakdown by model.