Audiogram Maker

An audiogram turns something people can only hear into something they can see in a feed. This one draws a new frame for every frame of video, so the waveform moves with the audio instead of sitting still behind it.

• Five animated styles. Bars, a filled line, a mirrored ribbon, a circle and a ring — each drawn from the audio at that instant.
• Three delivery sizes. Square for a feed, vertical for Reels, Shorts and TikTok, landscape for YouTube and a site embed.
• Cover art and a title. Layer artwork and a headline over the waveform, both inside the platform-safe area.
• Captions, written for you. Transcribed on your device in one press, or import an SRT or VTT — burned in and outlined so it reads over any background.
• A progress bar and a timer. The two cues that make a clip look finished rather than generated.
• Loudness targeting. Set a LUFS target and the clip is matched to it, with the peaks kept below clipping.
• Batch with one download. Render a queue of files on the same settings and take them all as a single ZIP.

The style decides how the clip reads at thumbnail size, which is where most people will first see it. Bar styles are the most legible; the radial styles work best when there is cover art to wrap.

• Bars. The clearest at small sizes and the safest default. Raise the bar count for a denser, more detailed look.
• Line. A filled ribbon around the centre. Calmer than bars and better suited to speech than to music.
• Mirrored. Bars growing from the centre in both directions — symmetrical, and the most decorative of the three linear styles.
• Circle. Bars radiating outwards. With cover art loaded the circle wraps it, which is the classic music-release look.
• Ring. The circle plus a progress sweep that fills as the clip plays, so the ring doubles as the progress indicator.

Each size reserves a margin the destination's own interface tends to cover, so nothing you place ends up underneath a caption or a row of buttons.

• Square. Works everywhere and crops least badly when a platform reframes it. A sound first choice if you post the same clip in several places.
• Vertical. The full-screen shape for Reels, Shorts and TikTok. The margins here are the largest because those apps overlay the top and bottom of the frame.
• Landscape. For YouTube and for embedding on a page of your own, where nothing is overlaid at all.

Short-form clips perform best at around 45 to 60 seconds across the major platforms, which is why the free tier renders up to 90 — long enough for a full pull-quote with room to spare.

Feeds play clips back to back, so a clip mastered quieter than the one before it reads as low quality before anyone has judged the content. The loudness control measures the clip and moves it to the target you set.

• Measurement happens after the trim, so a quiet pull-quote from a loud recording is judged on its own level rather than the whole file's.
• Fades are applied after measurement, so a long fade does not drag the measured level down and cause over-correction.
• Gain is capped so the loudest moment stays below full scale — a clip normalised into clipping sounds worse than the one you started with.
• Around -14 LUFS suits social platforms. -16 is the common podcast target. -23 is for broadcast delivery.

Most feed video is watched with the sound off, so captions are usually the difference between a clip that lands and one that is scrolled past. Bring an SRT or VTT and it is burned into the video.

• Cue times are shifted into the clip's own timeline, so a transcript for a whole episode captions a clip pulled from the middle of it correctly.
• Text is outlined rather than boxed, so it stays readable without punching a hole through the waveform behind it.
• A malformed cue is skipped and counted rather than rejecting the whole file.
• Automatic captions are generated on your device, in English. Your audio is never uploaded — the speech model is downloaded to you, about 40 MB on the first run, and cached afterwards. For other languages, import a transcript you already have.

Five steps, and the preview updates as you go so nothing is a surprise at the end.

1. Add audio. Drop an MP3, WAV, M4A, AAC, OGG or FLAC file. It is read on your device and never sent anywhere.
2. Start from a preset. Reel, Square, Podcast and Pulse each set a size, a style and a colour scheme in one press.
3. Trim to the moment. Set a start and a length, then scrub the preview to check the waveform lines up with what you want people to hear.
4. Layer what you need. Cover art, a title, captions, a progress bar, a timer — each is optional and each is previewed live.
5. Render. MP4 for posting, WebM for embedding. A batch comes back as one ZIP.

• Pick the moment first and build the design around it — a strong 30 seconds beats a well-styled 90.
• Keep one colour scheme across a series so clips are recognisable as yours in a feed.
• Add a short fade in and out when the clip starts or ends mid-sentence.
• Turn the progress bar on for anything over about 30 seconds so viewers can see how much is left.
• Check the vertical preview before rendering vertical — the safe margins are large for a reason.
• Raise the bar count for music and lower it for speech; speech has fewer peaks to show.

• Podcast pull-quotes. The strongest 40 seconds of an episode, captioned, as a feed post.
• Music teasers. A hook from a new track wrapped around the cover art.
• Voice notes and announcements. Turning something recorded on a phone into a post worth looking at.
• Series promotion. A batch of clips on identical settings, downloaded as one archive.

The audio, the artwork and the transcript stay on your machine — nothing is uploaded, and the video is built and encoded locally. That is also why there is no queue, no monthly cap and no watermark: there is no server bearing a cost that has to be recovered.

Powered by optimised processing algorithms.

• Podcast trimmer

  cut the exact segment before you render it

• Loudness meter

  check LUFS before publishing

• Audio converter

  convert a format this tool does not take

• Image to video

  build a clip from stills instead