Quick start
Three steps to the first answer:
- Sign up — 3 hours of every model open right away.
- Create a key under “API keys”. The secret is shown once.
- Set the address below and the key in your client.
https://tenzo-cloud.com/v1Authentication
Send the key in the Authorization header. For the Claude Code compatible mode the x-api-key header works too.
Authorization: Bearer tc-your-keyRequest examples
/v1/chat/completions and /v1/responses are supported, streaming and regular responses, tool calls and images for vision models.
curl https://tenzo-cloud.com/v1/chat/completions \
-H "Authorization: Bearer tc-your-key" \
-H "Content-Type: application/json" \
-d '{"model": "qwen-3.8-max", "stream": true,
"messages": [{"role": "user", "content": "Hello!"}]}'Editors and agents
Ready configs with your key and a check button are in the workspace under “Connect”.
export ANTHROPIC_BASE_URL=https://tenzo-cloud.com
export ANTHROPIC_AUTH_TOKEN=tc-your-key
export ANTHROPIC_MODEL=qwen-3.8-max
export ANTHROPIC_SMALL_FAST_MODEL=qwen-3.8-max
claudeModels
Put the model id in the model field. The list is also available at GET /v1/models.
| id | Name | Images | Context |
|---|---|---|---|
qwen-3.8-max | Qwen 3.8 Max | yes | — |
glm-5.3 | GLM 5.3 | no | — |
deepseek-v4-pro | DeepSeek V4 Pro | no | — |
glm-5.3-flash | GLM 5.3 Flash | yes | — |
deepseek-v4.1-flash | DeepSeek V4.1 Flash | yes | — |
Limits
- Tokens are unlimited.
- As many parallel requests as streams you bought; streams for one model stack.
- Up to 120 new requests per minute per account.
- Each key can be restricted by model, IP address, expiry and stream count.
Error codes
Errors follow the OpenAI format: an error object with message, type and code.
| code | HTTP | Meaning and what to do |
|---|---|---|
invalid_request | 400 | The request is malformed — check the body. |
invalid_api_key | 401 | The key is unknown, disabled or expired. |
no_access | 403 | No active subscription to the model, or the key doesn’t allow it. |
model_not_found | 404 | No such model — check the id. |
concurrency_limit | 429 | All your streams are busy or the per-minute limit is reached. Retry in a few seconds. |
upstream_error | 502 | The answer couldn’t be generated. Retry the request. |
model_unavailable | 503 | The model is temporarily unavailable. See the status page. |
{
"error": {
"message": "No active subscription for this model",
"type": "permission_error",
"code": "no_access"
}
}How we differ
What you get instead of paying per token.
Tokens
- Chat subscription
- Message caps, then waiting
- Pay-per-token API
- You pay for every token
Tenzo Cloud: No token limits
Price
- Chat subscription
- Flat, but with caps
- Pay-per-token API
- You find out from the monthly bill
Tenzo Cloud: Flat per period
Parallel work
- Chat subscription
- One chat window
- Pay-per-token API
- Limited by account tier
Tenzo Cloud: Buy as many streams as you need; streams stack
Request rate
- Chat subscription
- Not meant for automation
- Pay-per-token API
- Depends on the tier
Tenzo Cloud: Up to 120 requests per minute per account
Key control
- Chat subscription
- No keys
- Pay-per-token API
- Usually just on/off
Tenzo Cloud: Restrict by IP, model, expiry and streams
Web chat
- Chat subscription
- Yes
- Pay-per-token API
- No
Tenzo Cloud: Yes, on phone and computer
Payment
- Chat subscription
- Card only
- Pay-per-token API
- Card only
Tenzo Cloud: 5 ways: Kazakh and Russian cards, Visa/Mastercard, crypto, invoice
Transparency
- Chat subscription
- —
- Pay-per-token API
- Status page
Tenzo Cloud: Public model status page
API requests
- Chat subscription
- —
- Pay-per-token API
- Retained per the provider’s rules
Tenzo Cloud: We store neither requests nor responses
| Chat subscription | Pay-per-token API | Tenzo Cloud | |
|---|---|---|---|
| Tokens | Message caps, then waiting | You pay for every token | No token limits |
| Price | Flat, but with caps | You find out from the monthly bill | Flat per period |
| Parallel work | One chat window | Limited by account tier | Buy as many streams as you need; streams stack |
| Request rate | Not meant for automation | Depends on the tier | Up to 120 requests per minute per account |
| Key control | No keys | Usually just on/off | Restrict by IP, model, expiry and streams |
| Web chat | Yes | No | Yes, on phone and computer |
| Payment | Card only | Card only | 5 ways: Kazakh and Russian cards, Visa/Mastercard, crypto, invoice |
| Transparency | — | Status page | Public model status page |
| API requests | — | Retained per the provider’s rules | We store neither requests nor responses |
Privacy
The content of API requests and responses is written neither to the database nor to logs. Only technical metrics are kept: time, key, model, status, token count and duration.