Euqai Fusion API 2 Documentation

This is the documentation for Fusion 2; please use this for all new projects. For reference, the documentation for Fusion 1 is available here.

Welcome to the Euqai Fusion API. Fusion is not a single foundation model—it is a Sovereign Multi-Model Orchestration Engine. It provides state-of-the-art text generation, deterministic code execution, and visual analysis by routing queries through a multi-stage pipeline of specialized European AI clusters.

Designed as a drop-in replacement for OpenAI SDKs, Fusion 2.0 offers unique parameters for enhanced control, multi-model consensus verification, and automated tool self-healing.

Authentication#

All requests to the Euqai Fusion API must be authenticated. Include your API key in the Authorization header of your request.

Header Format: Authorization: Bearer YOUR_API_KEY

Endpoint: Chat Completions#

This is the primary endpoint for all interactions with the Fusion engine. Because we adhere to the OpenAI API contract, you can simply change your baseURL in existing SDKs to route through Fusion.

POST https://api.euqai.eu/v1/chat/completions

Request Body Parameters

Parameter	Type	Required	Description
messages	Array	Yes	A list of messages comprising the conversation so far.
model	String	Yes	Specify the engine pipeline. For the flagship orchestration, use euqai-fusion-v2.
stream	Boolean	No	If true, sends Server-Sent Events (SSE) for the fastest response time. Defaults to false.
response_format	Object	No	Use { "type": "json_object" } to guarantee a valid JSON response. Our internal robustJsonParser automatically strips markdown fences.
max_tokens	Integer	No	The maximum number of tokens to generate.
temperature	Float	No	Controls randomness (0.01 - 2.0). Lower values are more deterministic. Defaults to 0.7.
top_p	Float	No	Controls nucleus sampling. Defaults to 0.8.
stop	String / Array	No	Up to 4 sequences where the API will stop generating further tokens.
grounding	Boolean	No	Custom. If true, allows the Conductor to perform live web searches via the Brave API for real-time verification. Defaults to false.
language	String	No	Custom. A two-letter ISO 639-1 language code (e.g., "nl", "tr"). If omitted, the API automatically detects the language.
include_thinking	Boolean	No	Custom. If true, the internal ReAct scratchpad and model reasoning are prepended to the final response inside <think> tags. Defaults to false.

Under the Hood: The Fusion 2.0 Pipeline#

When you send a request, you aren't pinging a single LLM. You are triggering an autonomous pipeline:

1. ReAct Conductor: Assesses the query and determines if tools (Web Search, QuickJS Calculator, QuickJS Code Sandbox, or Page Fetcher) are needed.
2. Dual-Model Consensus: Two independent models generate factual answers. A comparator evaluates them. If they agree, the answer is verified.
3. Escalation: If the consensus diverges, or if the context exceeds 2500 tokens, the request automatically escalates to a 300-billion+ parameter flagship model for authoritative synthesis.

Quickstart Examples#

OpenAI SDK Drop-in Replacement (Node.js)

Because Fusion uses standard OpenAI formatting, you can use the official openai NPM package:

import OpenAI from "openai";

const openai = new OpenAI({
  apiKey: process.env.EUQAI_API_PRODUCTION,
  baseURL: "https://api.euqai.eu/v1" // Reroute to Fusion
});

async function main() {
  const completion = await openai.chat.completions.create({
    model: "euqai-fusion-v2",
    messages: [
      { role: "system", content: "You are a precise data extraction tool." },
      { role: "user", content: "Extract names and ages to JSON." }
    ],
    response_format: { type: "json_object" }
  });
  console.log(completion.choices[0].message.content);
}
main();

Multimodal: Image Input

You can send images for analysis by providing a content array with base64-encoded images.

{
  "model": "euqai-fusion-v2",
  "messages": [
    {
      "role": "user",
      "content": [
        {"type": "text", "text": "Describe this architectural diagram in detail."},
        {"type": "image_url", "image_url": {"url": "data:image/jpeg;base64,/9j/4AAQSkZJ..."}}
      ]
    }
  ]
}

Billing and Usage#

Euqai Fusion 2.0 provides $15/1M flagship-tier performance at baseline prices by dynamically routing easy queries to fast workers and saving heavy compute for edge cases.

• Text Input: €1.00 per 1M tokens
• Text Output: €4.50 per 1M tokens
• Grounding Cost: A flat 4000 prompt tokens is added only if the orchestrator executes a live web search.
• Image Input (Vision): Total Pixels / 1000 = Equivalent Prompt Tokens
• Image Output (Flux 2): Total Pixels / 100 = Equivalent Completion Tokens

The final, aggregated token counts are returned in the standard usage object, with an itemized breakdown in usage_details.