Product
Infistar, served at infistar.ai internationally and infistar.cc in mainland China.
About
About Infistar
Infistar is a unified multi-model API gateway: one API key and one base URL reach text, image, video, embedding, and reranking models from many vendors through OpenAI-, Anthropic Messages-, or Gemini GenerateContent-compatible formats, billed by actual usage.
At a glance
Service
Unified access, relay, management, billing, usage records, and API key management for AI model APIs. Some model capabilities come from third-party or upstream providers.
Pricing currency
Public prices are listed in CNY.
Integration
OpenAI, Anthropic Messages, and Gemini GenerateContent clients keep their own request format and switch only the base URL and key.
How pricing and settlement work
- Every public model has a listed price: text models per million input and output tokens, image and video models per successful result. The full list is on Models and Pricing.
- A request is pre-authorized and then settled by actual usage; every call in your account records the model, usage, and charge.
- One model can be served by several supply lines whose prices and availability differ; listed prices are the lowest currently available, and the account response at call time is authoritative.
Contact
Open source and license
The gateway and user console are a modified version of the open-source New API project, licensed under the GNU AGPL v3.0. https://github.com/QuantumNous/new-api
Frontend design and development by New API contributors.