Web URL Context Compressor with BM25 Budget Control

Choose this endpoint when an AI agent needs to ingest a live webpage but the raw HTML or full content would exceed the LLM context window or inflate token costs significantly. It is especially valuable in RAG pipelines, research agents, and competitive intelligence workflows where you need focused, token-efficient content. The BM25 budget mode is uniquely useful when you have a strict token ceiling and a specific query focus. Prefer this over generic web scrapers when token cost savings and ROI proof are important, or when you need structured/skeleton output rather than raw HTML.

• URL is unreachable or returns non-200 status — endpoint may return an error status
• URL points to a paywalled or bot-blocked page — distillation returns empty or partial content
• max_tokens is set below the minimum of 20 — validation error
• mode=budget used without a query parameter — BM25 focus may be undefined or fallback behavior
• Raw HTML passed in 'html' field is malformed — may produce garbled markdown
• Payment not included or insufficient — HTTP 402 response blocks the call
• Very large pages with huge raw_tokens may still exceed downstream LLM limits even after compression