One model, three modalities. Show it a picture, play it a sound, or just ask — thinkingmachines/Inkling handles all three and answers in text, with its reasoning shown as it works.
Sign in with your Hugging Face account to run the model. Requests go out on your Inference Providers allowance — nothing is stored by this Space.
Attach either an image or audio (up to 120s) — attaching one clears the other, because the provider can't take both in a single request yet.