Radient API Reference

Welcome to the Radient API documentation. This section provides detailed information about our API endpoints, authentication methods, request/response formats, and usage examples.

Overview

The Radient API allows developers to programmatically interact with Radient services, including:

• Managing your Radient Pass and credits.
• Accessing AI models and tools directly.
• Integrating Radient's capabilities into your own applications and workflows.

Authentication

Authentication with the Radient API is typically done via API keys. You will need to generate an API key from your Radient account dashboard.

Include your API key in the Authorization header as a Bearer token:

Authorization: Bearer YOUR_RADIENT_API_KEY

Certain public endpoints, like listing models or downloading agents from the marketplace, may not require authentication.

Endpoints

This section details the direct HTTP endpoints available through the Radient API. For higher-level client libraries and tool integrations, see the "Client Libraries & Tools" section.

(The Radient API is continuously evolving. More endpoints will be documented here as they become available.)

Chat Completions

These endpoints allow you to interact with chat-based language models.

POST /v1/chat/completions

Proxies OpenAI-compatible chat completions requests.

Request Body: OpenAIChatCompletionRequest (Refer to OpenAI documentation for standard fields. Radient-specific extensions or behaviors will be noted here.)

Responses:

• 200 OK: OpenAIChatCompletionResponse (Matches OpenAI's response structure).
  • id (string): A unique identifier for the chat completion.
  • object (string): The object type, typically chat.completion.
  • created (integer): The Unix timestamp (in seconds) of when the chat completion was created.
  • model (string): The model used for the chat completion.
  • choices (array): A list of chat completion choices. Each choice has:
    • index (integer): The index of the choice in the list.
    • message (object): The message generated by the model, with role and content.
    • finish_reason (string): The reason the model stopped generating tokens (e.g., stop, length, tool_calls).
  • usage (object, optional): Usage statistics for the request.
    • prompt_tokens (integer)
    • completion_tokens (integer)
    • total_tokens (integer)
• 400 Bad Request: Invalid request payload.
• 401 Unauthorized: API key is missing or invalid.
• 500 Internal Server Error: An error occurred on the server.

Example Request (Python using requests):

import requests
import json

api_key = "YOUR_RADIENT_API_KEY"
base_url = "https://api.radient.com/v1" # Replace with actual Radient API base URL

headers = {
    "Authorization": f"Bearer {api_key}",
    "Content-Type": "application/json"
}

payload = {
    "model": "gpt-4o",
    "messages": [
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello!"}
    ]
}

response = requests.post(f"{base_url}/chat/completions", headers=headers, json=payload)

if response.status_code == 200:
    print(response.json())
else:
    print(f"Error: {response.status_code}")
    print(response.text)

POST /v1/completions

Proxies OpenAI-compatible legacy completions requests. This endpoint is generally for older models or specific use cases. Prefer /chat/completions for newer models.

Request Body: OpenAICompletionRequest

Responses:

• 200 OK: OpenAICompletionResponse (Matches OpenAI's legacy completion response structure).
  • id (string)
  • object (string): Typically text_completion.
  • created (integer)
  • model (string)
  • choices (array): Each choice has:
    • text (string): The completion text.
    • index (integer)
    • logprobs (object, optional)
    • finish_reason (string)
  • usage (object, optional)
• 400 Bad Request, 401 Unauthorized, 500 Internal Server Error

Image Generation (via API Endpoint)

These endpoints allow direct interaction with image generation services.

POST /v1/images/generate

Generates images based on a prompt and other parameters. This endpoint supports both synchronous and asynchronous operations. For synchronous requests (sync_mode: true), the response will contain the image data directly if generation is quick. For asynchronous requests (sync_mode: false or if generation takes longer), the response will include a request_id which can be used with the /images/status endpoint to poll for results.

Request Body: ImageGenerationRequest

Responses:

• 200 OK: ImageGenerationResponse
  • request_id (string): Unique ID for the generation request.
  • status (string): Current status (e.g., "COMPLETED", "PROCESSING", "FAILED").
  • images (array of RadientImage, optional): List of generated images if status is "COMPLETED".
    • url (string): URL to the generated image.
    • width (integer, optional): Image width.
    • height (integer, optional): Image height.
  • provider (string, optional): The provider used for generation.
  • error (string, optional): Error message if generation failed.
• 400 Bad Request, 401 Unauthorized, 500 Internal Server Error

GET /v1/images/status

Retrieves the status and results of an image generation request.

Query Parameters:

Responses:

• 200 OK: ImageGenerationResponse (Same structure as the response from /images/generate).
• 400 Bad Request, 401 Unauthorized, 404 Not Found, 500 Internal Server Error

GET /v1/images/providers

Lists available image generation providers.

Responses:

• 200 OK: ProviderListResponse
  • providers (array of strings): List of provider IDs/names.
• 401 Unauthorized, 500 Internal Server Error

Web Search (via API Endpoint)

These endpoints allow direct interaction with web search services.

GET /v1/search

Performs a web search using the specified query and parameters.

Query Parameters:

Responses:

• 200 OK: WebSearchResponse
  • query (string): The original search query.
  • results (array of RadientSearchResult): List of search results.
    • title (string): Title of the search result.
    • url (string): URL of the search result.
    • content (string): Snippet or summary of the content.
    • raw_content (string, optional): Full raw content if include_raw was true and available.
  • provider (string, optional): The provider used for the search.
  • error (string, optional): Error message if the search failed.
• 400 Bad Request, 401 Unauthorized, 500 Internal Server Error

GET /v1/search/providers

Lists available web search providers.

Responses:

• 200 OK: ProviderListResponse
  • providers (array of strings): List of provider IDs/names.
• 401 Unauthorized, 500 Internal Server Error

Client Libraries & Tools (Python Example)

Radient provides client libraries and tool integrations to simplify interaction with the API. The following examples demonstrate usage with the Python local-operator library.

Image Generation Tool

The generate_image tool allows you to create images from text prompts.

Function Signature (Conceptual):

generate_image(prompt: str, image_size: str = "landscape_4_3", num_inference_steps: int = 28, seed: Optional[int] = None, guidance_scale: float = 5.0, num_images: int = 1) -> RadientImageGenerationResponse

Parameters:

• prompt (str): Text description of the image.
• image_size (str, optional): Size/aspect ratio. Defaults to "landscape_4_3".
  • Supported: "square_hd", "square", "portrait_4_3", "portrait_16_9", "landscape_4_3", "landscape_16_9".
• num_inference_steps (int, optional): Number of diffusion steps. Defaults to 28.
• seed (Optional[int], optional): Seed for reproducibility.
• guidance_scale (float, optional): How closely to follow the prompt. Defaults to 5.0.
• num_images (int, optional): Number of images to generate. Defaults to 1.

Returns: RadientImageGenerationResponse (see schema under /v1/images/generate endpoint).

Example (using local-operator tools):

# Assuming 'tools' is an initialized ToolRegistry from local_operator
# and radient_client is configured.

try:
    response = tools.generate_image(
        prompt="A futuristic cityscape at sunset, synthwave style",
        image_size="landscape_16_9",
        num_images=1
    )
    if response.images:
        for image in response.images:
            print(f"Generated image URL: {image.url}")
            # Code to download and save the image would go here
    else:
        print(f"Image generation status: {response.status}")
        if response.error:
            print(f"Error: {response.error}")

except RuntimeError as e:
    print(f"An error occurred: {e}")

Alter Image Tool

The generate_altered_image tool allows you to modify an existing image based on a text prompt.

Function Signature (Conceptual):

generate_altered_image(image_path: str, prompt: str, strength: float = 0.95, num_inference_steps: int = 40, seed: Optional[int] = None, guidance_scale: float = 7.5, num_images: int = 1) -> RadientImageGenerationResponse

Parameters:

• image_path (str): Path to the local image file to modify.
• prompt (str): Text description of how to modify the image.
• strength (float, optional): Strength of modification (0.0-1.0). Defaults to 0.95.
• num_inference_steps (int, optional): Number of inference steps. Defaults to 40.
• seed (Optional[int], optional): Seed for reproducibility.
• guidance_scale (float, optional): How closely to follow the prompt. Defaults to 7.5.
• num_images (int, optional): Number of images to generate. Defaults to 1.

Returns: RadientImageGenerationResponse (see schema under /v1/images/generate endpoint).

Example (using local-operator tools):

# Assuming 'tools' is an initialized ToolRegistry

try:
    # First, ensure you have an image file, e.g., 'input_image.jpg'
    response = tools.generate_altered_image(
        image_path="path/to/your/input_image.jpg",
        prompt="Make the cat wear a wizard hat",
        strength=0.8
    )
    if response.images:
        for image in response.images:
            print(f"Altered image URL: {image.url}")
            # Code to download and save the image
    else:
        print(f"Image alteration status: {response.status}")
        if response.error:
            print(f"Error: {response.error}")

except FileNotFoundError:
    print("Error: Input image not found.")
except RuntimeError as e:
    print(f"An error occurred: {e}")

Web Search Tool

The search_web tool allows you to perform internet searches.

Function Signature (Conceptual):

search_web(query: str, search_engine: str = "google", max_results: int = 20) -> RadientSearchResponse

Parameters:

• query (str): The search query.
• search_engine (str, optional): Search engine to use (e.g., "google", "bing"). Defaults to "google". (Provider dependent)
• max_results (int, optional): Maximum number of results. Defaults to 20.

Returns: RadientSearchResponse (see schema under /v1/search endpoint).

Example (using local-operator tools):

# Assuming 'tools' is an initialized ToolRegistry

try:
    response = tools.search_web(
        query="latest advancements in AI",
        max_results=5
    )
    print(f"Search Query: {response.query}")
    if response.results:
        for i, result in enumerate(response.results):
            print(f"\nResult {i+1}:")
            print(f"  Title: {result.title}")
            print(f"  URL: {result.url}")
            print(f"  Snippet: {result.content}")
    elif response.error:
        print(f"Search Error: {response.error}")
    else:
        print("No results found.")
except RuntimeError as e:
    print(f"An error occurred: {e}")

We are working diligently to bring you a robust and well-documented API. Please check back for updates.

If you have specific requirements or questions about upcoming API capabilities, feel free to contact us or email us.