Omnifact
One API. All models.

Omnifact AI Gateway

A drop-in replacement for programmatic LLM access. Build custom AI applications and internal tools against our AI Gateway using the models you trust, backed by the same enterprise data privacy, security, and GDPR compliance as the main Omnifact platform.

Read the documentation
chat completionscURL
curl https://connect.omnifact.ai/v1/gateway/chat/completions \
  -H "Authorization: Bearer $OMNIFACT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "eu/gpt-5.6-sol",
    "messages": [
      { "role": "user", "content": "Summarise this policy." }
    ],
    "stream": true
  }'

Point the base URL of any OpenAI-compatible SDK at connect.omnifact.ai/v1/gateway and keep the rest of your codebase unchanged.

One Endpoint. One Key.
One Invoice.

Unified Model Access

Access every model available in your Omnifact workspace—including EU-hosted deployments—through a single OpenAI-compatible endpoint. Switch models instantly by changing a single string parameter.

On-Demand European Routing

Prefix any model ID with eu/ to guarantee that traffic is routed exclusively through European infrastructure. Unprefixed IDs never route through EU backends.

Consolidated Enterprise Billing

API consumption runs directly on your existing Omnifact contract and appears on a single invoice alongside the rest of the platform. Prepay by invoice with no corporate credit cards or per-provider accounts.

Centralized Budget & Usage Control

Monitor spending and consumption across all providers in the Billing & Usage dashboard, managed under the same organization budget as Chat and Spaces.

Enterprise Data Governance,
Built Into Every Request

Every API request is governed by the same strict data privacy framework, data processing agreements (DPAs), and GDPR compliance that protect the main Omnifact platform.

  • Zero Model Training

    Prompts and completions are never used to train underlying provider models. Contractually guaranteed across every provider we route to.

  • Strict Regional Data Control

    Restrict API workloads to European infrastructure on a per-request basis, or enforce organization-wide policies that disable non-EU models entirely.

  • Zero Data Retention (ZDR)

    API payloads are never logged or stored by the AI Gateway. Select ZDR-compliant models for sensitive workloads requiring zero data persistence.

First-Party Support for
Our TypeScript AI Library

core-ai is a type-safe LLM abstraction library for TypeScript. With its dedicated Omnifact provider, there is no manual base-URL configuration or compatibility layer required—pass your organization key and start building with full type safety.

Learn moreGitHub
app.tsTypeScript
import { createOmnifact } from '@core-ai/omnifact';

const omnifact = createOmnifact({
  apiKey: process.env.OMNIFACT_API_KEY,
});

const model = omnifact.chatModel('eu/gpt-5.6-sol');
GPT-5.6 Sol
Claude Opus 5
Mistral Large 3
Gemini 3.7 Flash
Custom company model

Supported Models
Across Every Provider

Reach every model enabled in your Omnifact workspace—from OpenAI, Anthropic, Google, and Mistral to EU-hosted deployments and custom in-house models connected to your organization.

Fetch your organization's active model list dynamically via GET /v1/gateway/models.

Start Building with the Omnifact AI Gateway

Get an organization API key in seconds to connect your applications to top AI models with enterprise privacy and full GDPR compliance.

GDPR Compliant
ISO 27001 Certified
Hosted in Germany