About Takehound
Takehound is a self-serve video editor in the browser, with an agent that does the grunt work and learns how you cut, for people who talk to camera and are not editors.
I'm Kyle. I built Takehound. I have a full-time job, and I'm not an editor or a coach.
How it works
You add a long-form talk and your own clips, photos or slides. One press makes a first pass: pauses and retakes cut, your footage placed, your key lines on screen. The agent is Hound, and you steer him in plain words. You export an MP4, or a timeline (an FCPXML file) to finish in DaVinci Resolve.
Why it was made
The little time I get for videos should go to the part that makes a video mine, so I built Takehound to do the rest: the logging, finding and placing. That is the 80%. The story, the taste and the final cut are the 20%, and they stay with you.
The name
A newshound sniffs out the story. Takehound (sometimes searched as Take Hound) is the same idea for your own footage: for each line it can, it brings back the shot that fits.
What it is not
- Not a stock library. B-roll is only what you filmed.
- Not a video generator. Nothing is generated.
- Not a finished video from one prompt. You finish the cut in Takehound's editor (it exports an MP4) or DaVinci Resolve. It does not export to Premiere. Final Cut has not opened one.
- Not an editing service. It is a tool you or your editor run.
Other tools
Other tools work with your own clips too. AI B-roll from your own footage lists nine of them beside Takehound and says what each one does better than mine. I did not test those nine for that page, so each fact about them comes from the vendor's own pages.
Still choosing? Read how to choose an AI video editor for your own footage.
What it costs today
Sign in with Google for a free first video, up to 30 minutes, with no card. More videos use credits, and they never expire. Editing services' prices are in Video editor cost per video.
Your files
To describe and choose your shots, your browser sends small still images from your clips (not the video files), plus your script or transcript, the shot descriptions and the names of your clips and folders, through Takehound's server to Anthropic's API. The sound of a recording you add is transcribed on Takehound's own server, and a small copy of your cut goes through it to Google's Gemini model. The server keeps none of it, and nothing is used to train a model. The privacy page has the details.
How to try it
Takehound runs in a desktop browser: press Open the editor on the home page, or the coaches page if you coach, and sign in with Google. Or start from a free template.