API Overview

Conduit provides an OpenAI-compatible REST API that allows you to interact with various LLM providers through a consistent interface.

Base URL

The base URL for all API requests is:

http://your-conduit-host:5000/v1

Replace your-conduit-host with your Conduit server address (e.g., localhost for local development).

Authentication

All API requests require authentication using a virtual key in the Authorization header:

Authorization: Bearer condt_your_virtual_key

Virtual keys can be created and managed through the Conduit Web UI.

Core Endpoints

Conduit implements the following core endpoints:

Text Generation

Endpoint	Method	Description
/v1/chat/completions	POST	Create chat completions
/v1/completions	POST	Create text completions
/v1/embeddings	POST	Generate text embeddings
/v1/models	GET	List available models
/v1/images/generations	POST	Generate images (if configured)

Audio Services

Endpoint	Method	Description
/v1/audio/transcriptions	POST	Transcribe audio to text (Speech-to-Text)
/v1/audio/translations	POST	Transcribe and translate audio to English
/v1/audio/speech	POST	Generate speech from text (Text-to-Speech)
/v1/audio/speech/stream	POST	Stream generated speech
/v1/realtime/sessions	POST	Create real-time audio session
/v1/realtime/sessions/{id}/ws	WebSocket	Connect to real-time audio stream

Response Format

API responses follow the OpenAI format:

{
  "id": "cmpl-abcdef123456",
  "object": "chat.completion",
  "created": 1677858242,
  "model": "my-gpt4",
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "Hello! How can I help you today?"
      },
      "finish_reason": "stop",
      "index": 0
    }
  ],
  "usage": {
    "prompt_tokens": 10,
    "completion_tokens": 8,
    "total_tokens": 18
  }
}

Error Handling

Conduit returns standard HTTP status codes and error responses:

{
  "error": {
    "message": "Invalid request",
    "type": "invalid_request_error",
    "param": "model",
    "code": "model_not_found"
  }
}

Common status codes:

• 200: Success
• 400: Bad request (client error)
• 401: Unauthorized (invalid or missing API key)
• 403: Forbidden (permissions issue)
• 404: Not found
• 429: Too many requests (rate limit exceeded)
• 500: Server error

Common Parameters

These parameters are supported across multiple endpoints:

Parameter	Type	Description
model	string	The model to use (virtual model name)
user	string	User identifier for tracking and rate limiting
temperature	number	Controls randomness (0-2)
top_p	number	Controls diversity via nucleus sampling
n	integer	Number of completions to generate
stream	boolean	Stream responses as they're generated

Versioning

The API uses the /v1 prefix, aligned with the OpenAI API version. Future breaking changes may introduce new version prefixes.

Rate Limits

Rate limits are applied based on the virtual key settings. When exceeded, the API returns a 429 status code with a Retry-After header indicating when to retry.

Streaming Responses

For streaming endpoints, responses follow the Server-Sent Events (SSE) format, with each event containing a chunk of the response.

Next Steps

• Chat Completions API: Learn about the chat interface
• Embeddings API: Generate vector embeddings
• Models API: List and filter available models
• Audio API: Speech-to-Text, Text-to-Speech, and real-time audio