Frequently asked questions
Short answers to the questions we hear most. Each one links to the page with the full details.
The service#
What is Claudech?#
Claudech is an API for large language models. It accepts requests in the OpenAI Chat Completions format and the Anthropic Messages format, so existing SDKs and tools work by changing the base URL and key. There is also a web chat and a playground. See the Introduction.
Is Claudech affiliated with OpenAI or Anthropic?#
No. Claudech is an independent service. It implements their public request formats so your code keeps working; every request is served by a Claudech model.
Which models are available?#
The current list, with context windows, output limits and prices, is on Models and in GET /v1/models. Every model has a 1M-token context window and supports tool calling and image input.
Pricing and payments#
How much does it cost?#
Requests are billed per token at the published model rates: $0.50 per million input tokens, $1.90 per million output tokens and $0.010 per million cached input tokens. Every response reports its exact cost. See Tokens & billing.
How do I pay?#
Buy a token package, a custom amount or a monthly plan on the Pricing page. We accept PayPal, credit and debit cards (Visa, Mastercard, American Express and Discover, through PayPal, with no PayPal account needed) and cryptocurrency: BTC, ETH, USDT, USDC, DAI, BNB, POL, AVAX and TRX. Crypto payments need no identity verification. The networks for each coin are listed in Tokens & billing.
Do tokens expire?#
No. Purchased tokens never expire. Only the monthly allowance that comes with a subscription plan resets at the end of each billing period. See balances.
Can I get a refund?#
All purchases are final, and crypto payments cannot be reversed once they confirm on-chain. If you think a request was billed incorrectly, send its request_id to support and we will check the cost breakdown. See Tokens & billing.
Am I charged for failed requests?#
No. Requests that fail before the model produces output are never billed.
API keys and access#
How do I get an API key?#
Create an account, make your first purchase, then create a key under API keys. API access unlocks after the first purchase. See the Quickstart.
Which header do I send the key in?#
Either Authorization: Bearer sk-... or x-api-key: sk-.... Both work on every endpoint. See Authentication.
What are the rate limits?#
They depend on your plan, from 60 requests per minute on pay as you go upward. The table and response headers are on Rate limits.
Privacy#
Do you store my prompts?#
No. The content of API prompts and responses is not kept after the request finishes. The one exception is a 24-hour copy when your client sends an idempotency key, so a retried request returns the same result. See the Privacy Policy.
Compatibility#
Does it work with Cursor and Claude Code?#
Yes. Cursor uses the OpenAI-format base URL and Claude Code uses the Anthropic-format one. See Cursor setup and Claude Code setup, and all integrations for guides to Cline, Continue, Aider, Open WebUI, LangChain and more.
Which endpoints are supported?#
POST /v1/chat/completions, POST /v1/messages and GET /v1/models, plus account, usage and API key endpoints. Embeddings, images, audio, files and the OpenAI Responses API are not offered. See compatibility notes.
How do I move existing code over?#
Change the base URL, the API key and the model id. The migration guide covers both the OpenAI and the Anthropic SDKs.
Still have a question?#
Open a ticket from the Support page or email support@claudech.com.