Andrej Karpathy — Intro to LLMsY Combinator — How to get startup ideas
Your YouTube video
Paste a YouTube link. Nothing to upload.
←YouTube video
Make a competitive analysis
Turn YouTube videos into a competitive analysis
Reviews and comparisons name rivals out loud. Paste a few and get the matrix that falls out of them.
No link handy? Try one:
Free to try, no card. Your YouTube video opens in Ancher.
Caption track with timingsKey frames from the videoDescription and linked sources
Kept in parts
Every one of these stays separately addressable instead of collapsing into one block of text.
PlayersPositioningFeature matrixPricing
Competitive analysis
A side-by-side read on rivals: positioning, features, pricing, and the space nobody occupies.
Review and comparison videos are where products get named next to each other by people with no reason to be diplomatic. Ancher pulls every rival mentioned, the feature each creator actually tested on camera, the price shown on screen with the date it was shown, and the complaint that recurs across creators — then marks plainly which cells came from one video and are therefore one opinion.
The honest version
One creator's opinion is not a market
A single comparison video is an argument, not evidence. It tests the features that creator cares about, on their hardware, on the day they filmed. The analysis only becomes worth acting on when the same claim survives several creators who have no relationship to each other, and when the cells that did not survive are marked rather than quietly filled in.
One second of one video, at 11:42, read two ways.Illustrative. The figure is invented; the gap it falls into is not.
...and their pro tier is honestly not far off a month, which changes the maths.
Pricing page on screen — not readThe specific never reaches the draft
Competitive analysis
Pricing page on screen
$49 / seat / mo
from 11:42
Lands in the draft, with its timecode
- A cell is only filled when a feature was demonstrated on camera, not when it was named in passing.
- Prices are read off the screen with the video's date attached, because a pricing page from fourteen months ago is not pricing.
- Rivals named once are kept separate from rivals named across creators, so a one-off mention never reads as consensus.
What comes out
Competitive analysis, section by section
A side-by-side read on rivals: positioning, features, pricing, and the space nobody occupies.
Players
Every product named, split into the ones a creator actually opened and the ones only mentioned.
Positioning
How each rival is described in the creator's own words, which is usually blunter than the rival's own site.
Feature matrix
Feature by product, filled only where something was demonstrated on camera and left empty where it was not.
Pricing
Every price shown on screen, with the date of the video attached to it so staleness is visible.
Whitespace
Complaints that recur across creators and are answered by nobody in the matrix.
Videos-to-competitive-analysis prompt
The link on its own gets a summary. This is the instruction that produces the 5 sections above, in that order, with the rules that keep them honest. Paste it with your YouTube videos — in Ancher, or in whatever assistant you already use.
Build a competitive analysis from these videos. (1) List every product named, marking which were opened on camera and which were only mentioned. (2) For each, give the positioning in the creator's own words. (3) Build a feature matrix, filling a cell only where the feature was demonstrated on screen and leaving it empty otherwise. (4) Record every price shown on screen with the video's publish date beside it. (5) List complaints that recur across two or more creators and say which product, if any, answers each. Mark every claim that rests on a single video.
The detail, if you want it
How each part of a YouTube video becomes a section
| From the YouTube video | Into | Why |
|---|---|---|
| Caption track with timings | Players | Product names are spoken far more often than they are written on screen, so the caption track is where the roster comes from. |
| Key frames from the video | Feature matrix | A feature demonstration is a screen recording, so the evidence that a feature exists is a frame rather than a sentence. |
| Key frames from the video | Pricing | Pricing is shown, not read aloud, which is exactly the thing a caption-only tool returns nothing for. |
| Caption track with timings | Positioning | The sentence a creator uses to place a product is the positioning claim, and it is rarely the one the vendor writes. |
| Title, channel, duration, date | Whitespace | Publish dates order the complaints, so a gap that has stayed open for a year is distinguishable from one raised last week. |
Everything Ancher keeps from a YouTube video
Caption track with timings
Published and auto-generated captions both come through, and every line keeps the second it was spoken.
Key frames from the video
Frames are sampled and scored, so a number shown on a slide survives even though it was never said out loud.
Description and linked sources
The creator's own cited links come along, so a claim can be traced past the video.
Likes and comment counts
Reception is a weak signal about correctness and a strong one about which claims got attention.
Title, channel, duration, date
Basic provenance, kept so the finished work can attribute the video properly.
The four steps
- Add the YouTube video. Paste a YouTube URL. Ancher pulls the caption track with its timings, samples key frames from the video itself, and keeps the description links and engagement counts on the same source.
- Nothing gets flattened. Caption track with timings, key frames from the video, description and linked sources, likes and comment counts, title, channel, duration, date stay separately addressable instead of collapsing into one block of text.
- Ask for a competitive analysis. Ancher fills players, positioning, feature matrix, pricing, whitespace using the mapping above, not a generic summary pass.
- Check it before you use it. Every specific keeps a link back to where it came from in the YouTube video, so a claim can be verified without reopening the source.
Where this goes wrong
- Sponsored comparisons look identical to independent ones in a transcript. Check the description for a disclosure before letting a video decide a cell.
- Creators test what they use. A matrix built from three videos in the same niche will show the same four features and miss the rest of the category.
- A feature named but not demonstrated is a claim, not a finding. Those cells stay empty on purpose, and filling them in from a vendor site defeats the exercise.
Questions people ask
How many videos does this need to be useful?
Three or four from unrelated creators is where the matrix starts to mean something. Below that you are reading one person's setup and calling it a market, which is the most common way this kind of research goes wrong.
Why are some cells empty?
Because nobody demonstrated that feature on camera. An empty cell is a finding — it tells you the feature is either missing or so uninteresting that no reviewer bothered showing it — and inventing a value there would destroy the only thing the analysis is good for.
Is video better than reading vendor sites?
It is better for what breaks and what things cost in practice, and worse for completeness. Vendor sites list every feature; videos show you the four that people actually touch and the one that fails.
Bring your own YouTube videos
Everything you save lives in one workspace, so the competitive analysis is built from your sources — not from a model's memory of the internet.
Open Ancher →Not a YouTube video? The same competitive analysis also comes from X thread, LinkedIn post, Instagram post.