TL;DR: TokenProvider is a third-party API gateway: one OpenAI-compatible base URL gives you Claude, Claude Code, ChatGPT, and Gemini access with a single key. Tokens are metered per request; rates appear in the dashboard after sign-up. Not an official Anthropic, OpenAI, or Google product.
Unified AI Gateway for Cursor, Cline & Claude Code
One OpenAI-compatible endpoint for Claude, Claude Code, ChatGPT, and Gemini — a drop-in base URL for Cursor, Cline, and Claude Code. Pay-as-you-go per-request billing, sticky sessions, and instant setup.
Why developers route through TokenProvider
One reliable endpoint for Claude, ChatGPT, and Gemini — a drop-in base URL for your IDE, with pay-as-you-go pricing and no per-provider subscription to manage.
Unified API Gateway
Single OpenAI-compatible endpoint routes to Claude, Claude Code, ChatGPT, and Gemini. Swap models without rewriting code.
Smart Routing & Failover
Requests are load-balanced across upstream capacity with automatic failover, so you get higher effective rate limits and fewer 429s than a single key.
Transparent Pay-as-you-go
Pay-as-you-go, metered per request. See exact input/output cost for every call and export usage logs whenever you want.
Supported providers
Current model coverage — more arriving as upstream APIs open up.
How the gateway works
TokenProvider puts one OpenAI-compatible endpoint in front of Claude, Claude Code, ChatGPT, and Gemini. Point your client at the gateway, use your issued key, and requests are routed with sticky sessions so conversation context stays intact.
Because everything runs through one metered endpoint, you get a single dashboard for every model and pay-as-you-go pricing with no monthly commitment — deposit what you need, use it when you need it. For many workloads that works out cheaper than separate per-provider plans.
Works with any OpenAI-compatible SDK and as a drop-in base URL for Cursor, Cline, and Claude Code. Point your client at our endpoint, use your API key, and start shipping.
Guides, comparisons & code examples
Everything you need to evaluate and integrate the gateway. Start with a comparison page to pick the right model, then jump into a setup guide for your tool of choice.
Compare
Guides
Code examples
FAQ
How is pricing calculated?
Pricing depends on the current model and routing. Register for a free account to see live per-token rates — metered per request with no subscription, which for many workloads comes out below separate per-provider plans.
Does it work with Claude Code?
Yes. Set ANTHROPIC_BASE_URL to our gateway and use your issued API key. Sticky sessions keep long Claude Code conversations coherent.
Is it OpenAI-compatible for ChatGPT?
Yes. Drop-in replacement for api.openai.com — change the base URL, keep your existing code.
How does billing work?
Prepaid tokens. Every request debits your balance based on input + output tokens. Full per-request logs available.
Do you support Gemini and Claude Code together?
Yes — one account, one key, access Claude Opus/Sonnet, Claude Code, GPT-4o, and Gemini 2.0 through the same endpoint.
Start in 60 seconds
Create an account, top up a few dollars, get an API key, ship.