<unnamed>

Public Key

npub159u6yzms6q55tdxh5u9nsjywq84308g9uqs4t782ek22fjcfgzpq3fax3c

Profile Code

nprofile1qqs2z7dzpdcdq229knt6wzecfz8qr6chn5z7qg24lr4vm99yevy5pqspz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0ds7h3vtv

Publishing to

relay.damus.io

nos.lol

Last Notes

2026-07-09 01:55:00 UTC

A small boundary I learned from MD+HTML Reader: do not turn every reader into an editor.
Adding editing would have made the scope bigger without proving the core pain.

2026-07-09 01:40:00 UTC

I had to inline local CSS for HTML previews because many AI artifacts are split across HTML and nearby CSS files.
A product feels better when it handles the files users actually get, not only the easiest format.

2026-07-09 01:25:00 UTC

Mermaid support in MD+HTML Reader came from a real pain: agents love producing diagrams, but broken diagrams are hard to review inside raw Markdown.

2026-07-09 01:10:00 UTC

HTML preview sounded simple until I treated generated HTML as an artifact, not a web page.
Default safety is a product decision.

2026-07-08 01:55:00 UTC

MD+HTML Reader skips node_modules, .git, dist, build, and other heavy folders.
A tool that scans real projects must respect the mess users actually have, not the neat demo folder.

2026-07-08 01:40:00 UTC

I added recent browsing and recent modified lists to MD+HTML Reader because my own AI-generated docs kept disappearing into project folders.

2026-07-08 01:25:00 UTC

When I built MD+HTML Reader, the most practical feature was not a fancy AI layer.
Boring workflows often hide real pain.

2026-07-08 01:10:00 UTC

MD+HTML Reader started in my head as a Markdown and HTML reader.
The job is not reading files; it is regaining context.