~/deeprat ❯ sniff status --live
deeprat
❯ eight rats read the internet for you.
every source comes back in their teeth.
try
# not a replay. eight rats crawl the real web from your tab: wikipedia, arxiv, hacker news. ask them something and they go read it, then answer only from what they read.
8 rats awake · 0 pages · 0 links found · 0 hosts ·
❯ waking the rats…
-- press space for next page --
Arrow keys: Up and Down to move. Right to follow a link; Left to go back.
tail -f rat_01.log0 bitten
rat_01 sniffing
net
req
err
Tasks: 8 rats, 0 running
Read: 0 kB this session
Errors: 0
PIDRATSHOSTPAGESTIME+URL
man rat.md
- fetcheach rat asks /api/read for one page. it reads wikipedia, arxiv and hacker news through their public apis.
- readthe server pulls the real text and every link out of the page and sends it back to your tab.
- bitethe rat on screen walks the page and bites the links it likes. click a rat in the table to watch it.
- followit follows one bitten link and reads the next page. every hop, status code and byte count in the log is real.
- asktype a question. the rats search wikipedia, arxiv and hacker news, each one reads a page, and the answer is written only from those pages. every sentence carries the number of the page it came from.