MULTIMODAL PLAYGROUND

Evidence in. Decisions out.

Add your input, define the outputs, and run every field together.

Start with an example

01Input

Text + media

Drop an image, audio, or videoor choose files from your Mac

Add media

8 images · 1 audio · 1 video · 200 MB total

Media limits

Audio: up to 30 seconds. Silent video: up to 60 seconds; video with speech: 30 seconds. Video is sampled at a target of 1 frame/second, capped at 32 frames. Its soundtrack is included automatically. Choose one audio source. Longer inputs are rejected, not trimmed.

Shared instructions optional

02Define outputs

Each field has its own question and answer contract.

2 fieldsOne shared input · batched scoring