Process a single PDF page with OCR

Read a file and return the content in a format suitable for LLM consumption. Supported file types: - pdf: PDF documents (converted to im

Extract and read web page content from specified URLs using Tavily Extract. This tool extracts the main content from web pages, returning cleaned

Extract and read web page content from specified URLs using Tavily Extract. This tool extracts the main content from web pages, returning cl

Reset daily task selections (for testing)

Reset session tracking (for new decision/activity)

API root endpoint

Search the internet for information using Tavily (default, with AI-generated answers) or Jina AI. Tavily provides structured results with AI

Search the internet for information using Tavily AI search. Returns structured results with AI-generated answers. Args: query:

Set global state for MCP tools

Set global state for tools

Set list of hidden agent signatures

Start background tasks on startup

Submit completed work for evaluation and payment. Args: work_output: Your completed work as text (detailed answer to the task).

Submit completed work artifact for evaluation and payment. Args: artifact_path: Path to the completed work artifact file des

Test comprehensive analytics

Test daily state saving with completed tasks

Test that WorkEvaluator respects task-specific max_payment

Test fallback to default payment when values not available

Test that TaskManager loads and applies task values correctly

Test that task_values.jsonl has correct structure

Convert numpy/pandas types to JSON-serializable types

Execute MCP tool call

WebSocket endpoint for real-time updates