Overview
TOKI API documentation for integrating model services through OpenAI-compatible endpoints.
What is TOKI?
TOKI (元舟) provides model API services for developers. Its API shape follows common OpenAI API conventions: requests use Authorization: Bearer authentication, chat completions are sent to /v1/chat/completions, and responses follow the OpenAI Chat Completions structure.
If your application already uses the official OpenAI SDK or another OpenAI-compatible client, you can usually connect to TOKI by changing the baseURL and API key.
Key Features
- OpenAI-compatible API: Use standard OpenAI SDKs, HTTP headers, request bodies, and response shapes.
- API key management: Create, enable, disable, and manage API keys in the console, including supplier, quota, and expiration settings.
- Model directory: View available models, suppliers, context length, input and output types, and billing information.
- Token Plan: Purchase Token packages in the product center for supported model usage scenarios.
Getting Started
Quickstart
Get up and running in minutes with our quickstart guide.
Authentication
Learn how to authenticate your API requests.
API Reference
Explore the complete API reference documentation.
Models
Browse available models and their capabilities.
Token Plan
Learn about package purchase, quota usage, and usage rules.
Basic Flow
The typical integration flow is:
Register or sign in → Purchase or obtain quota → Create an API key → Call the API with an OpenAI-compatible clientExamples in this documentation use:
Base URL: https://www.tokiai.ai/v1
Endpoint: POST /chat/completionsAvailable models, quota, pricing, and rate limits are subject to what is shown in the TOKI console.