audiogram.video

You've recorded a great episode and want to share the best moment of it. Then you remember that Instagram and YouTube only play sound inside a video, so a podcast clip, an interview or a song needs a picture before you can post it. audiogram.video makes that video. You open a recording, choose the moment, and it draws a waveform that moves with the voice, your cover art and a title, plus captions that light up word by word as they're spoken.

The team behind it points to a simple problem. Platforms like Instagram, TikTok and LinkedIn don't play audio by itself, so podcasters need short video clips to promote episodes, and the usual audiogram tools are subscriptions with watermarks on their free plans.

Best 30 s or Best 60 s jumps to a lively stretch of talking, which saves you scrubbing through an hour of tape. One video can be up to 10 minutes long. There are 17 templates and nine wave styles, in square, vertical, 4:5 and widescreen shapes. Colours follow your cover art, and a look you like can be saved under My styles for the next episode.

Captions can be corrected by typing over them, and they come in a Highlight, Pop, Box or Plain style. You can also download them as SRT or VTT for another editor. The export is an MP4 at 1080p or 720p, at 30 or 60 frames per second.

Captions are the one step that leaves your device. With Automatic captions on, which is the default, the sound of the chosen clip is mixed to mono, compressed and sent through the site's server to DeepInfra, a US company running the Whisper speech-recognition model. When a recording first opens, that's up to 75 seconds of audio. The privacy page, updated November 9, 2026, says neither the site nor DeepInfra keeps the audio or the words, and your pictures, file name and the rest of the project aren't passed on.

Turn Automatic captions off and nothing is sent unless you press Transcribe. You can also type the captions yourself or open an SRT or VTT file. The team puts the trade-off the same way the privacy page does: with automatic captions switched off nothing leaves the device, but you then type the captions yourself.

There's no account, payment or watermark. The FAQ gives two limits: 10 minutes per video, and fair-use limits on captions. Per the privacy page, those are counted per connection using a one-way fingerprint of your IP address that's deleted after two days.

The example clips on the home page come from public-domain and Creative Commons podcasts, so you can try every control before opening a recording of your own. If you have an episode to promote this week, that's a good place to start.