BOZOSAGENT UTILITIES / SOURCE READER
01 / READ THE SOURCE

A webpage.
Just the useful part.

Public URL in. Readable Markdown out.

Keep the headings, links, and code. Get citation details and a fingerprint of what was retrieved. Ordinary parsing, with no AI model or sign-in.

NO AIFREE TO USEAPI + MCP
SOURCE URLPUBLIC HTTPS

We fetch the page and check its robots.txt. No cookies, page scripts, or stored content. The source site receives a request from BOZOS.

SUPPORTED

Public HTML, Markdown, and plain text.

FIRST RELEASE

Pages that require JavaScript, logins, and PDFs are outside this release. Sources have a 2 MiB size limit and a 15-second request limit.

EXTRACTED SOURCEAWAITING URL

Less page furniture.
More source material.

Your extracted Markdown will appear here, with the original source URL and retrieval details.

01
Read the main content
02
Keep links and structure
03
Take the citation with you
02 / CALL IT FROM YOUR AGENT

One source.
Three ways in.

Use this page, call the HTTP API, or connect the BOZOS MCP server and call readSource.

GET also accepts ?url=, max_characters, and format=markdown. POST keeps the source URL out of the incoming request path; hosting access logs may still exist.

curl -sS https://bozos.si/api/v1/source-reader \
+  -H 'Content-Type: application/json' \
+  -d '{"url":"https://example.com","max_characters":40000}'

JSON includes the Markdown, citation metadata, final URL, redirects, headings, links, retrieval time, hashes, and any truncation warnings. Read the API contract ↗