← Back to Discover

RollTok

RollTok lets applications call mainstream AI large models with monthly purchased quota and per-token billing.

Model APIs
Visit website ↗

What it does

RollTok gives applications access to multiple mainstream AI large models through one API Key. Users register with an email verification code, and a dedicated API Key is generated automatically. Next, they buy a monthly quota. The console provides the base_url and the API Key. Users choose either an OpenAI-compatible or a Claude-compatible interface and call models on demand. Charges are deducted in real time. The console shows balance and usage in real time. A bill is generated automatically each month. Supported providers include Qwen, DeepSeek, GLM, Kimi, Doubao and MiniMax. The price table on the documentation website lists the available models and their prices. Administrators configure that list dynamically.

Capabilities

  • The OpenAI-compatible endpoint is /v1/chat/completions, which fits the OpenAI SDK.
  • The Claude-compatible endpoint is /v1/messages, which fits the Anthropic SDK.
  • The Embedding endpoint is /v1/embeddings, which fits the OpenAI Embeddings SDK.
  • Listed models carry capability tags such as tool calling, structured output and deep thinking.
  • Listed models include GPT 5, Claude 4.8 Opus, Gemini 3.5 Flash, DeepSeek V4 Pro and Qwen3.8 Flash, with context lengths from 400K to 1M.

When to use it

  • You want to connect AI large models to your own application.
  • You use the OpenAI SDK or the Anthropic SDK and want to connect by changing only the client configuration.
  • You need an embeddings endpoint alongside chat models.