API GUIDE

How to turn a podcast episode into a video with an API

Audio has nowhere to go on YouTube, Reels or a feed without a picture attached. The usual free tools cap the duration, because rendering a video from an hour of audio costs the same whether anyone watches it or not. Here is the shape that works: render per clip rather than per episode, and put the words on screen.

The manual FFmpeg way

One filter graph for the waveform, the overlay geometry to work out by hand, and a second pass if you also want captions. Every parameter is a place to get it wrong, and you only find out after the encode.

ffmpeg -i episode.mp3 -loop 1 -i cover.jpg -filter_complex "[0:a]showwaves=s=1080x200:mode=cline:colors=white[wave];[1:v]scale=1080:1080[bg];[bg][wave]overlay=0:880" -shortest -pix_fmt yuv420p out.mp4

Do it with one API call

  1. 1

    Pick the clip, not the episode

    A full hour as video is the same encode cost whether it is watched or not. If you are repurposing, run find-highlights first and render only the moments you keep.

  2. 2

    Render the audiogram

    Send the audio as the input named "main" and, optionally, a cover image as "background". Without one you get a clean dark canvas. Choose the aspect ratio for where it is going: 1:1 for a feed, 9:16 for a Short or a Reel.

  3. 3

    Put the words on screen

    Run transcribe-video on the same audio. Its second output is an SRT, and its download URL goes straight into add-subtitles-to-video on the audiogram. A clip in a feed is mostly watched on mute, so the waveform alone gives a viewer nothing to read.

REST
curl https://kinopipe.com/api/v1/tools/audiogram \
  -H "Authorization: Bearer kp_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "inputs": [
      { "id": "main", "url": "https://example.com/episode.mp3" },
      { "id": "background", "url": "https://example.com/cover.jpg" }
    ],
    "options": { "aspect_ratio": "1:1" }
  }'
Or just ask your agent
Connect the KinoPipe MCP server and say it in plain language:

Turn this podcast clip into a square video with my cover art behind the waveform, and burn the captions in.

Connect MCP

FAQ

Do I need a background image?

No. Without one the waveform is drawn over a solid dark canvas, which is enough for a feed. A cover image mostly helps recognition when the clip is shared.

Can I get the waveform on its own?

Yes. generate-audio-waveform draws it as a single image, a transparent PNG by default with a configurable size and colour, if you are composing the video somewhere else.

Which aspect ratio should I use?

Square reads well in a timeline and wastes the least space next to text. Vertical is for Shorts, Reels and TikTok. Rendering the same clip twice is cheap, so it is worth doing both rather than guessing.