Clean, readable transcripts for TED and YouTube videos
Problem
People who absorb text far better than audio/video (a common trait in the ADHD community) have no good way to read TED or YouTube content. TED transcripts require multiple clicks plus copy-paste and come out broken mid-clause, while YouTube auto-captions are low quality and unreadable as text.
Opportunity
A web tool or browser extension that extracts, cleans and properly formats transcripts (punctuation, paragraphs, speaker labels) from popular video sites, with an ASR fallback pipeline for videos with bad or missing captions.
Market analysis
The pain is real for text-first readers, but extraction itself is already a solved, saturated market of free tools, and pasting a URL into any LLM chatbot produces cleaned text today. The only defensible angles are reading workflow (paragraphing, speaker labels, export to note apps) and coverage of caption-less videos via ASR, neither of which commands obvious pricing power.
Market · Text-preferring readers, ADHD and accessibility communities, students and researchers who mine talks for quotes; steady but low-urgency demand.
Pricing · Extraction tools are mostly free or freemium; meeting-transcription SaaS anchors willingness to pay around $7-19/month (Otter Pro, Fireflies Pro/Business), which is a stretch for a single-purpose reader utility.
Pros
- + Genuinely felt pain with a simple, cheap MVP (caption fetch plus LLM cleanup pass).
- + ASR fallback for caption-less videos is a real coverage gap most free tools skip.
- + Browser-extension distribution matches the exact moment of frustration.
Cons
- − Free incumbents everywhere (youtubetotranscript.com, YTTools, youtube-transcript.io) with zero switching costs.
- − Any LLM chatbot already cleans a pasted transcript for free.
- − Site markup changes on TED/YouTube mean a permanent scraping-maintenance treadmill.
- − Weak willingness to pay for a one-shot utility.
Existing / similar tools
Source
Hacker News (Ask HN)
The hard truth is that extraction is the commodity part and cleaning is one LLM call, so the moat has to live somewhere else entirely: workflow and coverage. A winner here looks like an extension that turns any video page into a properly typeset article (paragraphs, speaker labels, keyboard-friendly reading mode) and syncs highlights into Notion/Obsidian, not another paste-a-URL site. The ASR fallback for videos with bad or missing captions is the one technically meaningful differentiator, since free tools mostly depend on existing captions. Non-obvious risk: TED and YouTube markup changes will silently break scrapers, so expect the maintenance burden, not feature work, to be the main cost of staying alive.