AI API FOR DEVELOPERS

One API. Every model.

Denizen Blu speaks the OpenAI-compatible dialect you already know. Point your base URL at us and keep your code.

Authentication

Every request carries a bearer token. Create keys from your dashboard and store them as environment variables—never in client code.

Shell
export DENIZEN_API_KEY="blu_sk_..."

Your first request

A single POST returns a completion. Conduit picks the execution path; you just name the model.

cURL
curl https://api.denizenblu.com/v1/chat/completions \
  -H "Authorization: Bearer $DENIZEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "blu-sovereign-70b",
    "messages": [{ "role": "user", "content": "Hello, world." }]
  }'

SDKs

Official clients wrap auth, retries and streaming for you.

TypeScript
import { DenizenBlu } from "@denizenblu/sdk";

const blu = new DenizenBlu({
  apiKey: process.env.DENIZEN_API_KEY,
});

const res = await blu.chat.completions.create({
  model: "blu-sovereign-70b",
  messages: [
    { role: "user", content: "Draft a kind welcome note." },
  ],
});

console.log(res.choices[0].message.content);
Python
import os
from denizenblu import DenizenBlu

blu = DenizenBlu(api_key=os.environ["DENIZEN_API_KEY"])

stream = blu.chat.completions.create(
    model="blu-sovereign-70b",
    messages=[
        {"role": "user", "content": "Explain Conduit."},
    ],
    stream=True,
)

for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")

Endpoints

The full surface is OpenAI-compatible, plus a workloads API for managed compute.

  • POST/v1/chat/completionsChat with any catalog model. Streaming supported.
  • POST/v1/completionsClassic text completion for legacy prompts.
  • POST/v1/embeddingsVectorize text for search and RAG.
  • GET/v1/modelsList every model available to your plan.
  • POST/v1/workloadsSize and reserve managed compute for a job.

Rate & fair use

No per-token billing. Plans include generous fair-use capacity measured as telemetry, not invoices. Organizations can reserve dedicated throughput via the workloads API or an enterprise plan. Explore enterprise plans