Hearsay reads a published podcast transcript — your own show, or anyone else’s — finds a moment worth reacting to, and renders a finished vertical video about it.
Commentary written in a persona you pick: the angle it takes, and how it argues. Then read aloud in a synthetic voice you pick separately. Burned-in captions. Generated cover artwork. And a short excerpt of the original audio, sitting under the point as the evidence for it.
Roughly sixty seconds long. Then it writes a caption for each platform and publishes to your connected accounts — after you’ve watched the thing and said yes.
The videos are openly machine-made. Every one carries a line saying so, and every caption credits the source podcast and the speaker by name.
The disclosure and the credit are part of the video itself. Nothing Hearsay makes pretends a person recorded it.
How it works
Four steps, and two of them are you
-
Pick a moment
Hearsay reads the published transcript and surfaces the moments worth reacting to. You choose the one you want.
-
It writes and voices the commentary
Written in the persona you chose, then narrated in the synthetic voice you chose.
-
It renders the video
Vertical, with captions burned in, cover artwork generated, and the excerpt of source audio that backs the point up.
-
You review, then publish
You watch it. If it's good, Hearsay writes the caption for each platform and posts it to your connected accounts.
A human approves every video before it posts
Nothing publishes on its own. There is no unattended mode.
What comes out
A finished video, and the copy to go with it
- Format
- 1080 × 1920, vertical. Around sixty seconds long.
- Captions
- Burned into the frame.
- Your two picks
- A persona, which decides what the commentary argues and how it sounds on the page. And a voice, which reads it aloud. They are chosen separately.
- Evidence
- A short excerpt of the original podcast audio, so the point being made has the actual moment underneath it.
- Attribution
- The source podcast and the speaker, named on screen and again in the caption.
- Disclosure
- A line on every video saying it was made by a machine.
- Artwork
- Cover art generated for the video.
- Copy
- A caption written for each surface it is going to — YouTube, Instagram, Facebook, TikTok and X — rather than one caption pasted five times.
Early access
Put your name down
Hearsay is not open yet. The first round of accounts goes to this list, in the order people joined it.
One thing worth knowing before you sign up: Hearsay runs on your own API keys — Google Gemini for the writing and the artwork, ElevenLabs for the voice. Generation is billed to you by those providers directly. Nothing passes through us.
Who it's for
Two jobs, one tool
Podcasters clipping their own show
You publish an episode and the good bits stay buried in it. Hearsay reads your transcript and turns those moments into short videos, with your show and your speakers named on every one.
Commentators reacting to someone else's
You already have opinions about other people's episodes. Hearsay writes the commentary in the persona you've chosen and puts the quoted moment in the video, credited to whoever actually said it.