API Documentation

JuntoRouter is fully compatible with the OpenAI API format. Use your API key to get started.

1. Get Your API Key

Your API key looks like: sk-junto-xxxxxxxxxxxx

Keep your key safe. It will only be shown once when created.

2. Make Your First Request

junto uses the same format as the OpenAI API. Just change the base URL and use your junto API key.

curl https://juntorouter-api.moonshine-studio.net/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-4o-mini",
    "messages": [
      {"role": "user", "content": "Hello!"}
    ]
  }'

3. Use with OpenAI SDK

Python

from openai import OpenAI

client = OpenAI(
    api_key="YOUR_API_KEY",
    base_url="https://juntorouter-api.moonshine-studio.net/api/v1"
)

response = client.chat.completions.create(
    model="openai/gpt-4o-mini",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)

Node.js / TypeScript

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: "YOUR_API_KEY",
  baseURL: "https://juntorouter-api.moonshine-studio.net/api/v1",
});

const response = await client.chat.completions.create({
  model: "anthropic/claude-sonnet-4-5",
  messages: [{ role: "user", content: "Hello!" }],
});
console.log(response.choices[0].message.content);

4. Available Models

Browse all models at Models. Use the model slug as the model parameter.

openai/gpt-4o

openai/gpt-4o-mini

anthropic/claude-sonnet-4-5

anthropic/claude-haiku-4-5

google/gemini-2.0-flash

google/gemini-2.0-pro

5. Dynamic Variants

Append a suffix to any model to control routing behavior:

:nitroRoute to the fastest provider (highest throughput)

:floorRoute to the cheapest provider (lowest price)

:onlineAugment prompt with web search results

:thinkingEnable extended reasoning (Anthropic models)

// Use the fastest provider for GPT-4o
{ "model": "openai/gpt-4o:nitro" }

// Use the cheapest provider
{ "model": "openai/gpt-4o:floor" }

6. Streaming

Add stream: true to get Server-Sent Events (SSE):

curl https://juntorouter-api.moonshine-studio.net/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "anthropic/claude-sonnet-4-5",
    "messages": [{"role": "user", "content": "Tell me a story"}],
    "stream": true
  }'

7. Provider Preferences

Control routing with the provider object:

{
  "model": "openai/gpt-4o",
  "messages": [...],
  "provider": {
    "order": ["anthropic", "openai"],
    "allow_fallbacks": true,
    "data_collection": "deny",
    "require_parameters": true
  }
}

order — Provider priority order

allow_fallbacks — Allow fallback to other providers

data_collection — "deny" = only no-data-collection providers

zdr — true = only Zero Data Retention providers

require_parameters — Only route to providers supporting all request params

8. Media Generation

JuntoRouter supports image, audio, and video generation through dedicated endpoints.

Image Generation

curl https://juntorouter-api.moonshine-studio.net/api/v1/images/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/gpt-image-1",
    "prompt": "A futuristic city skyline at sunset",
    "size": "1024x1024",
    "n": 1
  }'

Parameters: model (required), prompt (required), size, n, quality, style

Audio Speech (TTS)

curl https://juntorouter-api.moonshine-studio.net/api/v1/audio/speech \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "openai/tts-1",
    "input": "Hello, welcome to JuntoRouter!",
    "voice": "alloy"
  }' --output speech.mp3

Parameters: model (required), input (required), voice (alloy, echo, fable, onyx, nova, shimmer), speed, response_format

Video Generation

curl https://juntorouter-api.moonshine-studio.net/api/v1/video/generations \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/veo-3.1-fast-generate-preview",
    "prompt": "A cat playing piano in a jazz club",
    "duration": 6
  }'

Parameters: model (required), prompt (required), duration (seconds), size. Video generation may take 1-5 minutes.

Media Response Format

{
  "data": [
    {
      "b64_json": "iVBORw0KGgo...",  // base64-encoded media
      "url": "https://...",            // or a download URL
      "content_type": "image/png"
    }
  ],
  "model": "openai/gpt-image-1",
  "cost": 0.042
}

Responses include a data array with b64_json or url, plus a cost field in USD. Audio speech returns binary audio directly (not JSON).