Documentation · Start
Your first request
Five steps from a new account to a reply and its receipt. You need an account and somewhere to run Python, Node or curl.
Step 1: Sign up
Use Sign up on tenvor.io. No card is needed, and a new account starts with $0.25 of free credit.
Then sign in to the Console. It is where you create keys and read your usage.
Step 2: Create an API key
In the Console, open API keys and choose Create key. Tenvor may ask you to verify it is you again before it creates, replaces or revokes a key.
The key starts with sk-dc-. It is shown once, and Tenvor cannot show it again. Copy it into a password manager or your secret store before you close the page.
Your account has one key. If you lose it or it leaks, replace it: the old key stops working at once, and the new one is shown once.
Treat it like a password
A key can spend your balance. Keep it out of source code, URLs, screenshots and chat messages.
Step 3: Set your key and change the base URL
Put the key in an environment variable named TENVOR_API_KEY. Your code reads it from there, so the key never has to be written into a file.
Then point your OpenAI client at Tenvor. You keep the SDK. You change the base URL, the key and the model id.
- Base URL
https://orchestrator.tenvor.io/v1- Key
sk-dc-…, read fromTENVOR_API_KEY- Model id
chat-35b-moe
Step 4: Send one request
Install the SDK first: pip install openai for Python, or npm install openai for Node. Then run the sample. The highlighted line is the one that sends your request to Tenvor.
Prefer streaming for long replies. Without streaming the server waits up to 4,100 seconds, then allows 10 seconds to write the result. If delivery breaks, fetch the stored result with X-Request-Id or the same Idempotency-Key.
Step 5: Read the response and its receipt
The reply is in choices[0].message.content. Beside it are two blocks of numbers. usage counts tokens (roughly, pieces of words). billing is the itemised receipt for this request.
{
"id": "chatcmpl-…",
"object": "chat.completion",
"model": "chat-35b-moe",
"choices": [
{
"index": 0,
"message": {
"role": "assistant",
"content": "Hello! How can I help you today?"
},
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 14,
"completion_tokens": 9,
"total_tokens": 23
},
"billing": {
"full_input_tokens": 14,
"cached_input_tokens": 0,
"output_tokens": 9,
"charged_micro_usd": …
}
}full_input_tokens | 14 | Tokens you sent, at the full input rate. |
|---|---|---|
cached_input_tokens | 0 | Input tokens billed at the cached rate, half the full input rate. It reads 0 on a first request and on any request that did not ask for reuse. |
output_tokens | 9 | Tokens in the reply. |
charged_micro_usd | … | What you are charged, in millionths of a dollar: 1,000,000 is $1. The rates are on Pricing. |
You pay per token, for what goes in and what comes out. For each request you pay the larger of the tokens you used or the minimum per request. The rates and the minimum are on Pricing.
Your SDK may hide fields it does not know about. To see the whole body, use curl, or ask your SDK for the raw response. The Usage page in the Console lists the same charge.
Streaming changes where the receipt is
A streamed response ends with token usage if you ask for it. Its itemised receipt is at GET /v1/results/{id}, where the id is the X-Request-Id response header.
Where to go next
Pi
Our recommended coding tool. One command sets up the model.
OpenCode
Use Tenvor from your terminal, with approvals on.
JetBrains AI Assistant
Add Tenvor as a provider in your IDE.
Chat Completions
Streaming, tools, results and every error.
Billing and receipts
What a request costs, what is held, and how to read a receipt.
Content checked . Live limits and prices can change.