RAG Diet – BM25 Budget-Controlled URL/HTML Distillation

Choose this endpoint when you need to compress a specific webpage or raw HTML into a strict token budget before feeding it to an LLM, especially when BM25-relevance-ranked extraction matters (i.e. you have a focus query). Prefer it over a generic web scraper when token cost and ROI measurement are important, or when you need mode flexibility (Markdown, structured JSON, skeleton, or budget). It's ideal for RAG pipelines where context window space is limited and you need provable compression metrics.

• URL is unreachable or returns non-200 status — distillation fails with an error status
• max_tokens set too low (below 20) — rejected by schema validation
• BM25 query provided without mode=budget — query may be ignored or trigger unexpected behavior
• Raw HTML payload is malformed or empty — distillation returns minimal or empty output
• Payment (x402 HTTP 402) fails or USDC balance is insufficient — request blocked before processing
• Rate limiting or server overload causing slow or failed fetch of target URL