API Gateway

​

What is the API Gateway?

• One key, multiple providers: Access models from OpenAI, Anthropic, Google, Mistral AI, and more with a single API key.
• Drop-in replacement: Switch between providers and models with minimal code changes using existing OpenAI-compatible SDKs.
• Unified billing & observability: Monitor your spending and usage across all providers in one place. No credit card needed for separate providers.
• Runs on your existing budget: Consumes your organization’s existing AI budget, billed through your existing Omnifact contract.

​

Getting Started

1. Prerequisites: Ensure that API Gateway access has been enabled for your organization by Omnifact. (Note: If this feature is not enabled for your organization, any requests to the API will return a 403 Forbidden error.)
2. Opt-in: A team administrator must enable the Gateway in the Developer Tools section.
3. Authentication: All requests to the API Gateway must include your API key (which can be generated in Team Settings > Developer Tools) in the x-api-key header (or as a Bearer token if using OpenAI SDKs).

​

Available Models

​

GET /v1/gateway/models

curl https://connect.omnifact.ai/v1/gateway/models \
  -H "Authorization: Bearer YOUR_API_KEY"

{
  "object": "list",
  "data": [
    {
      "id": "gpt-4o",
      "object": "model",
      "created": 1715368132,
      "owned_by": "openai"
    },
    {
      "id": "claude-4-sonnet",
      "object": "model",
      "created": 1718841600,
      "owned_by": "anthropic"
    }
  ]
}

​

EU-Hosted Models

• The /v1/gateway/models endpoint may return both standard model IDs (e.g., gpt-4o) and EU-specific model IDs (e.g., eu/gpt-4o).
• To enforce routing to an EU backend, prefix the model ID with eu/ in your request. Standard model IDs will route to non-EU backends only.

​

Chat Completions

​

POST /v1/gateway/chat/completions

import OpenAI from 'openai';

const client = new OpenAI({
  baseURL: 'https://connect.omnifact.ai/v1/gateway',
  apiKey: 'YOUR_API_KEY', // typically process.env.OMNIFACT_API_KEY
});

async function main() {
  const response = await client.chat.completions.create({
    model: 'gpt-4o',
    messages: [{ role: 'user', content: 'Hello, how are you?' }],
    stream: true,
  });

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

main();

from openai import OpenAI

client = OpenAI(
    base_url="https://connect.omnifact.ai/v1/gateway",
    api_key="YOUR_API_KEY"
)

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {"role": "user", "content": "Hello, how are you?"}
    ],
    stream=True
)

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

curl https://connect.omnifact.ai/v1/gateway/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{
    "model": "gpt-4o",
    "messages": [
      {
        "role": "user",
        "content": "Hello, how are you?"
      }
    ]
  }'

​

Request and Response Constraints

• Messages: The last message in the array must have the user role. The system or developer roles are only permitted at the start of the message array. The developer role is supported.
• Content: Message content can be a string or an array of { type: "text", text: "..." } objects (arrays are flattened).
• Limits: A maximum of 1024 messages per request, and up to 200,000 characters per message.
• Parameters: Supported parameters include temperature (0–2, scaled per provider; some models may reject it), max_completion_tokens (or max_tokens), stream, and stream_options.include_usage.
• Response Model Field: The response returns the canonical model type, which may not necessarily match the requested ID (e.g., if you requested an eu/... prefixed model or an alias).

​

Usage & Billing

​

Error Codes

• 401 Unauthorized: Missing or invalid API key.
• 402 Payment Required: No active subscription (subscription_required) or credits exhausted (insufficient_quota). An active subscription is validated on every chat completion request.
• 403 Forbidden: API Gateway access is not enabled for your organization.
• 404 Not Found: The requested model was not found or is disabled.
• 429 Too Many Requests: Upstream provider rate limit reached.

​

FAQ