Scribe

Internal tool
$ ~/aspaloyy/scribe

Research system · internal tool
The tool I built because the research for Manas outgrew note-taking.


What it does

01

Scribe decodes any source of information into structured markdown. Not a summariser — a decoder: whatever the format, what comes out is clean, structured text that can be worked with.

Sources it reads

02
  • YouTube videos
  • Articles and long-form pages
  • Posts and threads on X
  • Any web page, from a bare link
  • Social media content
  • Almost any file format, PDFs included
  • Spoken audio, transcribed via Whisper / Omnilingual

What happens next

03

From there the material is broken down semi-automatically into finished, self-contained modules and filed into a knowledge base. The point is that a subject researched once stays usable: it does not decay back into forty open tabs and a folder of half-read PDFs.

Why it exists

04

Manas needs real research at volume — canon, lexicon, historical detail, variant versions of the epic, academic commentary. Manual note-taking does not survive that scale. Scribe was built to make the research tractable, and it now handles everything I read for the project.

Status

05
Functionally it works and I use it every day. What remains is packaging: turning it into a product that does not depend on one specific machine, including a Telegram bot as the mobile entry point. It is an internal tool for now, not a public product — I would rather say that than overstate it.