Execute the tool with given parameters Args: params: Tool input parameters Returns:

Monospace text for code snippets.

Load configuration from JSON file

Run agent with streaming output Args: user_message: User's input message max_iterations: Maximum num

Create a deep copy of the current rating system.

Create configuration from environment variables

Start a new chess game. Returns: Dictionary with initial board state

Run the learning agent to complete the task. Args: max_steps: Maximum steps for browser-use agent

Search for documents matching the query Returns list of (doc_id, score, debug_info) tuples

Make a move on the chess board. Args: move_str: Move in UCI (e.g., 'e2e4') or SAN (e.g., 'Nf3') format Returns:

Validate and resolve file path. Args: file_path: Path to the file Returns: Resolved Path object

Create configuration from environment variables

Load configuration from JSON file

Execute a task using available tools Args: task: The task to execute max_iterations: Maximum number

Run the interactive demo

For now, we approximate the cost of a datapoint as the cost of the input (output tokens are priced as input tokens as well).

Generate speech from text prompts. Orpheus supports emotion/expression tags in your prompts: - <laugh>, <giggles>, <

Convert chunk to text format for indexing

Update ratings after a match between two models. Args: model_a: Identifier for first model model_b:

Create config from environment variables

Process a user request with optional tool usage (OpenRouter format) Args: user_request: The user's request or qu

执行搜索并生成答案 Args: user_question: 用户问题 max_iterations: 最大搜索迭代次数（防止无限循环） Returns:

Draw a curved arrow using a quadratic bezier.

Extract transcript from a YouTube video. Args: video_id: YouTube video ID or URL language_code: Language code for the tr

Create config from environment variables

Get memory as a formatted string for LLM context

Save configuration to JSON file

Check if path is within allowed directories.

Resolve path relative to workspace.

Run a shell command and return results with proper process management.

Write additional files to tmp directory.

Execute an action and return (feedback, reward, done).

Execute a task using ReAct pattern with standard OpenAI tool calling Args: task: The task to execute

Execute a tool by name with given arguments

Get list of available actions in current state.

Process a user query using the ReAct pattern. Args: user_query: The user's question stream: Whether

Get a natural language description of the current game state.

Load a chess position from FEN notation. Args: fen_string: FEN string representing the board state Returns:

Search for chemical compounds in PubChem database. Args: query: Search term or identifier search_type: Type of search (n

Start the vLLM server Args: wait_for_ready: Wait for server to be ready timeout: Maximum time to wai

Extract a single piece of content to text

Check if path is within workspace.

Pretty print test result Args: result: Test result dictionary

Resolve path relative to current directory.

Apply color to text if colors are enabled

Execute a task using available tools with system hints Args: task: The task to execute max_iteration

获取 API Key Args: api_key: 可选的 API key，如果提供则使用，否则从环境变量获取 Returns: API key

Get content as bytes

Get financial statements for a company. Args: symbol: Stock ticker symbol statement_type: Type of statement (income_stat

Get current leaderboard sorted by rating. Returns: List of tuples (model, rating, matches, wins) sorted by ratin

Get a specific test case by ID. Args: test_id: The test case ID Returns: TestCa

Non-agentic RAG mode: Simple retrieval + LLM response. Args: user_query: The user's question stream:

Convert chunk to text format for indexing

Write content to a file with safety checks. Args: path: File path (relative to workspace or absolute)

Get the current board state in a structured format.

Add a new memory card. Args: card: The memory card to add Returns: Memory ID in

Small rounded badge/tag.

Have a conversation with the user Args: message: User message Returns: Assistan

Evaluate a single test case Args: test_id: The test case ID to evaluate Returns:

Find files matching a pattern (similar to Unix find command) Args: pattern: File name pattern (supports wildcard

Get API key for the specified provider Args: provider: Provider name (defaults to configured provider)

Get content as base64 encoded string

Get the current game status. Returns: Dictionary with game status information

Get comprehensive index statistics

Get current stock quote information. Args: symbol: Stock ticker symbol (e.g., AAPL, MSFT) Returns: TextCont

Generate system hint message to append to context

Get OpenAI-compatible tool schemas

A dashed group boundary with a label at top-left.

Index a new document and return its ID

Load test cases from YAML files Args: category: Optional category filter (layer1, layer2, layer3)

打印彩色文本

Print a formatted header

Process a query and return the final answer as a string Args: query: The query to process max_iterat

Process recent conversations and update memories Returns: Processing results with list of operations

Non-agentic RAG mode: Simple retrieval + LLM response. Args: user_query: The user's question stream:

Reset the game to initial state.

Search for relevant experiences. Args: query: The search query top_k: Number of top results to retur

'Train' the agent through in-context learning over multiple episodes. Note: Unlike traditional RL, there's no explicit training - jus

Validate the configuration Returns: True if configuration is valid, False otherwise

Horizontal stage pipeline with an optional dashed feedback loop.

Save timer state to storage.

Add conversation chunks to the index Args: chunks: List of conversation chunks to index rebuild: Whe

Add conversation chunks to the index Args: chunks: List of conversation chunks to index rebuild: Whe

Add a message to conversation history