Docs

Use the API

Bluusun speaks the OpenAI chat completions API. Point an existing SDK at this base URL and send a model id from the catalog.

Base URL

https://www.bluusun.com/api/v1

Authorization is Bearer bsu_…. Create and revoke keys in the dashboard. The key is shown once.

Chat completion

curl https://www.bluusun.com/api/v1/chat/completions \
  -H "Authorization: Bearer bsu_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.3-flash",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Python and Node use the official OpenAI SDK with base_url set to the URL above. stream: true returns server-sent events. Tool calls go through when the upstream route supports them.

If you omit max_tokens, the gateway caps the completion at 4,096 tokens or at whatever your balance can still pay, whichever is smaller.

Models and balance

• GET /v1/models lists the live catalog and the per-million rates.
• GET /v1/balance returns balance_usd for the key’s account.

Billing

A successful response is billed as fresh input tokens × input rate, plus cached input tokens × cache-read rate, plus output tokens × output rate. Rates are USD per million tokens, the same numbers as the models page. The charge is rounded up to the nearest millionth of a dollar.

Responses include X-Bluusun-Price-USD and X-Bluusun-Balance-USD. The same figures land in the dashboard usage log. HTTP errors are not billed.

Add credit with Stripe from the dashboard. Packs are $10, $25, $50, and $100. Credits do not expire. There is no subscription.

Errors

• 401 missing, unknown, or revoked key.
• 402 the balance cannot cover this prompt.
• 404 unknown model id.
• 503 no upstream route under the published price. Nothing is billed.

Request bodies are limited to 1 MB.