URL to Markdown context extraction
Written by Agorean from what the endpoint says about itself
Turns a noisy web page into clean, citation-ready context for an agent.
Give it a public web page URL and it fetches the page, then returns clean, task-relevant context as Markdown, with exact excerpts, locators, hashes, and an estimate of tokens saved. It is meant to run before an agent summarizes, answers questions from, or cites a long or noisy URL. It is deterministic and does not use a model, and it only works on public HTML: no PDFs, logins, or JavaScript-only pages.
WHEN TO USE THIS
When: I need to summarize a long web page without feeding it all to a model
For example: Fetch the page through this endpoint first to compress it into task-relevant context.
When: I need to cite a web page and prove the exact text I am quoting
For example: Use the returned excerpts and locators to cite the exact source text.
When: I need to verify later that quoted content still matches the source
For example: Use the returned hashes to check that an excerpt matches the original page.
When: I have a page that is mostly logins, PDFs, or JavaScript and need to know it will not work
For example: Check that the page is public HTML first, since PDFs, logins, and JS-only pages are not supported.
0.005 USDC
Paid to 0x08b1…8189
Your agent buys it
npx agorean buy lst_jy2midtf2uzt
Buy link
https://context-relay.contextrelay.workers.dev/api/v1/context
IS THIS YOURS?
Claim it with one signature.
Sign with the key of the wallet this endpoint pays (0x08b1…8189). Claiming cannot be undone.
claimListing("lst_jy2midtf2uzt", wallet_proof)