Benchmarks
What famous docs pages cost an agent, raw vs indexed.
Every row below is a real page read through the Lyrenth index, with the token economics reported by the API itself: what the raw HTML would cost a model to ingest, what the clean AIDocument costs instead, and the difference. Measured 2026-07-05. Original data, reproducible with one API call.
| Page | Raw HTML tokens | AIDocument tokens | Smaller | Saved |
|---|---|---|---|---|
| Stripe API reference Payments | 307,902 | 2,000 | 154x | 99.4% |
| Vercel Functions docs Serverless | 237,390 | 2,149 | 110x | 99.1% |
| Cloudflare Workers docs Serverless | 94,963 | 1,377 | 69x | 98.5% |
| GitHub REST API quickstart Developer platform | 88,614 | 4,876 | 18x | 94.5% |
| Kubernetes Pods concepts Infrastructure | 131,977 | 7,475 | 18x | 94.3% |
| React useState reference Frontend | 111,042 | 8,156 | 14x | 92.7% |
| MDN: Overview of HTTP Web fundamentals | 56,692 | 5,085 | 11x | 91.0% |
| Python asyncio docs Languages | 6,175 | 1,251 | 4.9x | 79.7% |
| Node.js fs module Languages | 259,563 | 102,372 | 2.5x | 60.6% |
| PostgreSQL SELECT reference Databases | 29,509 | 17,500 | 1.7x | 40.7% |
The spread is honest: dense reference pages like PostgreSQL's SELECT save less, documentation apps wrapped in heavy chrome save the most. Cost basis: $3.00 per 1M input tokens (Claude Sonnet / GPT-4 Turbo class), as reported in the API's economics block.
Get API access →Free tier · no credit card