Handle chat completions using the Anthropic API. Args: messages: List of message dictionaries with 'role' and 'content'

Handle chat completions using the SambaNova API. Args: messages: List of message dictionaries with 'role' and 'content'

Check if API key exists for service

Check if Ollama is running and has the required model.

Clean up streaming for a task.

Clean up temporary files.

Create a client based on configuration.

Create a client based on the provider type.

Create or retrieve a cached client instance for the specified provider.

Create local and remote client pair.

Create a new directory or ensure it exists. Args: path: Path to the directory to create Returns: Success message or error

Create FastAPI endpoint for metrics.

Create and return a new minion instance with configured clients.

Create an enhanced image prompt based on chapter content and story theme. Args: chapter_content: The text content of the chapter

Extract PII from the given text and return as a dictionary. Args: text (str): The text to extract PII from Retu

Extract text from a PDF, TXT, Python, or Markdown file.

Extract text from all supported files in a folder.

Format the output of a tool

Formats the summaries into a string for the remote client, organized by search query Args: query_results: Dictio

Generate images for each chapter

Get the extended agent card for authenticated users.

Get contents of a specific file in the workspace. Args: file_path: Path to the file relative to workspace direct

Get detailed file/directory metadata. Args: path: Path to the file or directory Returns: Dictionary with file metadata

Return the base64-encoded SHA256 hash of an ECDSA public key in DER format. :param public_key: A cryptography public key object :return:

Get security schemes for the agent card.

Initialize the voice generator model.

Initialize a client based on provider name.

Initialize the local and remote clients outside of the run_protocol function.

Check if we're in mobile mode (based on session state toggle).

List directory contents with [FILE] or [DIR] prefixes. Args: path: Path to directory (defaults to current directory) Returns:

Open a URL in the specified browser or default browser.

Parse a model string in the format provider/model_name.

Process a folder containing text, PDF, and image files. Returns: Tuple containing: - Concatenated text content from all text and

Profile a model, with optional manual quantization parameters. Args: name: The model name from the model_to_params table

Read the contents of a file. Args: file_path: Path to the file to read Returns: Contents of the file as a string

Read the contents of multiple files at once. Args: file_paths: List of paths to the files to read Returns: Dictionary mappin

Render the Deep Research UI component. Args: minions_instance: An initialized DeepResearchMinions instance

Dependency to require specific scopes.

Revoke an API key.

Revoke an OAuth2 client.

Run tests in the workspace. Args: test_commands: List of test commands to run Returns:

Sanitize credentials in action objects for safe display in logs and UI. Returns a copy with passwords masked.

Save uploaded file to a temporary directory and return the path.

Recursively search for files/directories. Args: path: Starting directory for search pattern: Search pattern (glob format)

Send a message to the supervisor and get a response

Send a message to the supervisor and get a streaming response

Set up a test environment with some sample files.

Show the conversation history.

Show the conversation history.

Convert TextEmbedding to dictionary representation. :returns Dict representation of the TextEmbedding

Context manager to track PDF processing.

Update client pool statistics.

Validate secure endpoint by checking if it uses HTTPS and is reachable.

Wait for server to be ready.

Create new file or overwrite existing file. Args: file_path: Path to the file to write content: Content to write to the file

Initialize the synchronous MCP client

Initialize Ollama Client. Args: model_name: The Ollama model to use (default: "llama-3.2") temperature: Sampl

Authenticate with Hugging Face using the provided token or environment variable.

Handle multiple messages using the /batch endpoint for parallel processing.

Build and return the Apriel model and processor. Apriel models use AutoModelForImageTextToText and AutoProcessor. Returns:

Build and return the model and tokenizer from the checkpoint path. Supports HF Hub models, local paths, and PEFT adapters. R

Handle chat using native Cohere SDK v2. Args: messages: List of message dictionaries with 'role' and 'content' k

Handle chat using OpenAI compatibility API. Args: messages: List of message dictionaries with 'role' and 'conten

Handle chat completions using the native xAI SDK with file support. Args: messages: List of message dictionaries with 'r

Handle chat completions using the OpenAI-compatible API. Args: messages: List of message dictionaries with 'role' and 'c

Handle chat using the websearch agent through the conversations API. Args: messages: List of message dictionaries with '

Check if required dependencies are installed.

Check if required dependencies are installed.

Remove old completed tasks based on retention policy.

Conduct final integration review of the completed project.

Convert MCP tools to Ollama tools format.

Convert standard message format to Apriel format with content as list of dicts. Text-only version. Args:

Create File Search tool for RAG capabilities.

Create Google Search tool for web search capabilities.

Create a sandboxed namespace for code execution.

Create URL context tool for retrieving content from URLs.

Create a websearch agent with the specified configuration.

Decide whether to use local or remote model for the current turn based on cost sensitivity.

Generate default metadata for clients if not provided. Returns: Dictionary mapping client names to default metadata in A

Auto-detect appropriate test commands based on project structure. Returns: List of test commands to try

Detect if there are URLs in the message content. Args: messages: List of message dictionaries