[BRLTTY] Firelynx v0.1.0

[Nicolas Pitre]

Hi all,

That thing I teased is out. Firelynx bridges lynx and Firefox: a small 
local proxy runs your lynx requests through a headless Firefox, so 
JavaScript actually executes and modern sites render, then hands clean, 
semantic HTML back to lynx. You keep lynx's fast, braille-friendly 
interface — your keybindings and workflow — while getting a real shot at 
JavaScript-heavy sites that normally serve a blank page. Reddit, for 
instance, works (it even clears Reddit's JavaScript challenge). It all 
runs on your own machine; HTTPS, forms and logins, and basic handling 
for JavaScript dialogs and multi-factor auth may sometimes work too.

Honesty up front: this is alpha. It's slower than plain lynx (a full 
Firefox renders every page), content extraction sometimes trims too 
much, and the one thing I really wanted to crack — Google Search — still 
trips Google's bot detection and lands on a CAPTCHA. I tried a lot of 
approaches and couldn't beat it, so I'm not holding the release back any 
longer. DuckDuckGo is the default and works well; other engines are 
hit-or-miss (Bing sometimes stalls on a cookie-consent page). For 
everyday browsing plain lynx is still more reliable; reach for Firelynx 
when a site really needs JavaScript.

Linux only; needs Firefox, lynx, and Python/Selenium. It's a git clone 
plus ./install.sh — Fedora and Debian/Ubuntu steps are in the README:

  https://github.com/npitre/firelynx

Enjoy !

Nicolas