Chat completions

POST /v1/chat/completions — request shape, streaming, and code samples in curl, JS, and Python.

Request#

POST /v1/chat/completions with the standard OpenAI body.

curl https://api.darkroom.computer/v1/chat/completions \
  -H "Authorization: Bearer $DARKROOM_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "darkroom/noir",
    "messages": [{"role": "user", "content": "Explain the exploit class in this contract."}],
    "stream": true
  }'

JavaScript (OpenAI SDK)#

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.darkroom.computer/v1",
  apiKey: process.env.DARKROOM_API_KEY,
});

const stream = await client.chat.completions.create({
  model: "darkroom/noir",
  messages: [{ role: "user", content: "Write me a brief on X." }],
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? "");
}

Python#

from openai import OpenAI

client = OpenAI(
    base_url="https://api.darkroom.computer/v1",
    api_key=os.environ["DARKROOM_API_KEY"],
)

resp = client.chat.completions.create(
    model="darkroom/noir",
    messages=[{"role": "user", "content": "Write me a brief on X."}],
)
print(resp.choices[0].message.content)

Parameters#

Field	Notes
`model`	One of the Darkroom models, e.g. `darkroom/noir`.
`messages`	Standard `role` / `content` array. `darkroom/vision` accepts image content parts.
`stream`	`true` streams tokens as server-sent events; omit for a single buffered reply.
`temperature`, `max_tokens`, …	Standard sampling parameters are passed through.

Billing#

Calls meter against your wallet balance on input + output tokens, identical to the room.

← Previous

Overview

Models endpoint