Token Spark is an AI model gateway: one API endpoint and one key give you access to frontier models from multiple providers. The API is OpenAI-compatible — point your base_url at https://tokenspark.uno/v1 and keep your existing code.
Frontier models from OpenAI, Anthropic, Google, DeepSeek, Moonshot, Alibaba, Zhipu AI and Tencent Hunyuan are available through a single key, spanning the GPT, Claude, Gemini, DeepSeek, Kimi, Qwen, GLM and Hunyuan families. Usage is billed per token, and every call is itemized.
Beyond model access, the platform also provides art data and annotation services for AI systems, plus ready-to-use agent apps.
What is Token Spark?
Token Spark is an AI model gateway (API aggregator): one API endpoint and one key for frontier models from multiple providers, OpenAI-compatible, billed per token with itemized statements.
How do I connect to Token Spark?
Register, create a key in the console, and set your base_url to https://tokenspark.uno/v1. See the docs center for the full walkthrough.
Which models are supported?
Frontier models from OpenAI, Anthropic, Google, DeepSeek, Moonshot, Alibaba, Zhipu AI and Tencent Hunyuan. The current list is shown on the models page and in the console.
How does billing work?
Pay per token with no subscription fee. Model, token usage and amount for every call are itemized in the console; prices are shown there in real time.