Sofya

Sofya wraps four research operations into a single streamable HTTP endpoint: search returns full page content instead of snippets, fetch converts URLs into clean markdown including PDFs and DOCX files, extract uses AI to pull structured data from pages, and research runs multi-source synthesis with citations. Each operation costs between 1 and 25 credits depending on depth. The markdown output is designed for agent consumption, and the research endpoint breaks queries into sub-questions before synthesizing results. Reach for this when your agent needs more than search snippets or when you want document fetching, extraction, and synthesis through one API instead of chaining multiple tools.