Sound
Generate a sound effect from a short text description and drop it onto your timeline at the playhead, as a new audio layer in After Effects.
Sound turns a short text description into a ready-to-use sound effect and places it directly on your timeline. Describe what you want to hear — a cinematic whoosh, a punchy impact, a soft UI click — and Prism generates the audio and adds it as a new layer at the playhead. It is the fastest way to sketch in a sound without leaving After Effects or hunting through a library.
How to use it
Sound places audio into the composition you are working in, so make sure one is open and your playhead is where you want the sound to begin.
In the Prism panel, switch to the Tools tab and choose Sound.
Type a short description — focus on texture, material, and motion. For example: "cinematic whoosh transition, punchy". The more specific the description, the closer the result.
Confirm to generate. Prism creates the sound effect and drops it onto your timeline as a new layer.
You can also ask your AI to do this for you. If you say something like "add a punchy whoosh here," it can run Sound on your behalf and place the result on the timeline, using the same description-to-audio flow.
What you get
- A new audio layer. The generated sound is imported and added as a fresh layer in your active composition. Nothing existing is changed or overwritten — your other layers are untouched.
- Placed at the playhead. By default the new layer starts at the current time indicator, so the sound lands exactly where you expect it.
- Automatic length. You do not set a duration. The length is inferred from your description, which keeps short, punchy effects short and lets longer textures breathe. Generated effects are capped at 30 seconds.
- One tidy undo. Placing the sound is a single undo step, so you can remove it in one action if it is not the right fit.
Notes and limits
- Write the prompt like a brief: name the texture, material, and motion ("metallic clang, reverb tail" rather than just "clang").
- A composition must be open before you run it — the active comp is the context. If none is open, Prism will ask you to open one first.
- Generation runs on the Prism server. If Prism can't reach it, see troubleshooting for connection debugging.
- If a result is not quite right, run it again with a more specific description — each run is a fresh generation.
- Sound costs 2 credits per second of audio, with a 10-credit minimum per generation. For plan caps, metering, and the full per-tool cost table, see usage and limits.
Related
- Subtitles — the other AI-native audio tool
- Usage and limits — credit costs and monthly caps
- The MCP panel — where the Tools tab lives
- The active comp is the context