Soundbite
The best minute of your Space, as a video for any feed. Open the recording, drag over the moment, and the words are written out and timed as they're said. Pick a look and download an MP4.
How to use it
- Open the recording. Press Open audio at the top, or drop the file on the editor. An MP3 or a WAV can be hours long; an M4A works too, up to an hour or so. Only the sound comes in: the picture starts as a blank white page.
- Pick the moment. Drag across the waveform to select it, then drag the green edges to trim. Drag along the thin strip just above the waveform to move the playhead. On a long recording, use the Zoom buttons above the waveform; a small map of the whole recording then appears at the top of the timeline, and you can click or drag on it to move along. A recording opens fully selected, so the clip can be the whole of it or any stretch you trim it to, however long. To keep two moments, put the playhead between them and press Split, then trim each half. To add another file, press Add recording, or drop it on the timeline: it comes in after the clip as a part of its own. The numbered parts under the timeline play back to back; press one to trim it, or drag it into a new order.
- Check the words. Captions start as soon as the recording opens and work through all of it. Whatever you select is captioned next, and the line beside Auto captions at the top says how far they've got. Open the Captions tab to fix any word. If a clip's captions came out wrong, press Redo clip captions there.
- Make it yours. Start from one of sixteen modern podcast designs under Design, or build on the white page; each design keeps its own background. Elements has a section for every kind of part, shown in your design's colours: captions and sound, shapes, lines, photo frames and grids, free photos, moving backgrounds, charts, tables, device mockups and 3D shapes, with a search and the ones you used last at the top. Drop a photo on a frame to fill it. Press anything on the picture to select it, drag to move it, pull a handle to size it or the round one to turn it, and double-click words to type them; new words hug their box as you type. Shift-click or drag a box to pick several, and right-click for the full menu. The bar above the picture sets the font, size, colour (gold and silver included), spacing and effects, crops and frames pictures, and opens Animate, where each tile shows its movement and Preview plays it once. Start over clears the design, and Undo brings it back. All of it but the music is remembered for next time.
- Download the MP4. Post it to X, TikTok, Instagram, YouTube or LinkedIn. The captions are part of the picture, so they show with the sound off.
Good to know
- The recording, the words and your pictures stay in this browser. Nothing is uploaded.
- The speech model, 44 to 66 MB, comes from Hugging Face the first time and is kept by the browser after that.
- An hour of audio takes about 5 to 20 minutes to caption on a laptop, longer on a phone. Whatever you select is captioned first, so you can download that moment while the rest is still being written.
- Square suits the feed, 4:5 fills more of a phone, 9:16 fills it for TikTok, Reels and Shorts, and 16:9 fits YouTube or LinkedIn.
- The Music tab has four free tracks written for Soundbite. Any other music should be yours to share: platforms take down or mute clips with songs they don't have the rights to.
- Download offers four sizes. 360p and 720p are free and carry a small "made with drk.vc/engines/soundbite" label in the corner; 1080p and 1440p, marked with a crown, come without it after a short ad, one for 1080p and two for 1440p.
- Soundbite is free and open source: the code is on GitHub.