Everything in the catalog

Generates a video with a soundtrack from image, audio and video references using MiniMax-H3.

Chat with Gemma 4 models fully in the browser via MediaPipe

Image to high-res 3D mesh with Hi3Dv1 refinement

OCR, VQA, Thinking and Object Detection.

Applies the mixing style of a reference track to your own track.

Generates a full song with vocals from a text description using the MiniMax Music 3 model.

Translates Spanish text into Nahuatl.

Identifies the font used in an image of text.

Plain chat interface for Gemma 4 31B instruction-tuned.

Generate and edit speech audio with one unified speech model.

Bilingual text-to-speech that supports designing a voice, cloning one and directing how the line is delivered.

Runs the Ternary-Bonsai-2-27B language model on your own device through WebGPU, with no server needed.

Generates background music for a video and adds it to the clip.

Writes descriptive captions for an image in styles suited to Stable Diffusion 3 prompts.

NVIDIA's OCR model extracting text from images.

Expands a text-to-image prompt into several variations with MagicPrompt.

Draw a line and the model continues the drawing as you sketch.

Generates song ideas from text, with an optional melody as a starting point.

Discriminator of Chest Vocie and Falsetto

Krea 2 Turbo: text-to-image + identity edit

One-pass calibrated decisions on images

Generate images in 8-steps

Unleash the Power of Visual Grounding

Process and analyze Arabic text with several language tools.