Invideo for Youtube

Search inside Youtube videos (think Ctrl-F)

Invideo is the first chrome extension enabling you to search inside of Youtube videos.

For the techies out there, it also finds keywords in the video using NLP.

Comment if you like it and don't hesitate to share it with friends! I know they'll love this just as much as you will! :)

Would you recommend this product?
5 Reviews4.0/5
Hey there! I originally built Invideo to help me save time while watching Youtube videos, but what's the fun in being the only one to experience this? So I thought I'd share the love by sharing Invideo with you guys/gals :) Y'all are the first to experience this, so your thoughts matter a lot to me. Let me know what you think in the comments below! I promise I'll get to you.
Nice tool Victor!
Thanks for the support @vladkorobov!
nice job Victor :)
Thanks for the support @mrtampham!
I think it was 2012, when i first thought about this problem. At the time my idea was to make a DB of movie subtitles and after getting some search results, reference the timestamp in that movie. But doing this only for Youtube is a lot of straignforward for referencing the actual video. Great work guys. I have yet to try it out, but good start. Do you have any plans on making some kind of DB for Close Caption of Videos (auto-generated or manually added), so that you can search in a lot of video in one go. An interesting application is for listening to some word or phrase prnounciation, can be useful of people learning the language of the video. If you have some meta info about the spoken words in the video like, who is saying those words, or maybe Person A is saying this and Person B is saying this, then this can be used for Data digging purposes. Specially for Politicians which BTW can be useful for Shows like Late Night with SC or The Daily Show.
Hey @vinit_agrawal1! Yeah absolutely, is there any particular application you'd find useful? I think the one you mentioned may be a bit difficult, but getting transcripts in bulk should be doable. Try out actually. (200k videos). This is another project I hacked together with 3 other friends. We built the initial dataset with an aggregator at a Hackathon instead. It's going to be pretty slow since I'm running it on a free Heroku server and it's not active, but it should work. I'm currently not storing any transcripts through my extension. I think a headless scraper works better.
@victor_cheng The one application i am thinking right now for this, is to use this data set for training a model, whose purpose is to understand and model human conversations. I think with just the audio transcribed data set, it is not enough. Meta data like who is saying those words would be quite helpful. But even with this data set, of video subtitles and video images... You know what, i am no fuckin cloe what i am talking about. I don't even know how am i gonna use this dataset and how is it gonna help me in the approach towards building a good conversational model of machines. I mean we as humans think about so many things while talking about something, and we just expect the machines to randomly shuffle some symbols and start to make sense. That is idiotic. Sorry for the Rant.
@vinit_agrawal1 I can definitely see how the data could be useful (i.e. creating a better experience for bots) and will take this into account in my next release :).Thanks for the feedback!
This is awesome! Incredibly useful and great UX!