Everything in the catalog

Recursive self-improvement of LLM agent harnesses

A Multi-Modal World Model for Reconstruction

Multimodal 2B model that runs locally in the browser on WebGPU.

Official WebLLM demo that loads a model into the browser and answers prompts without a backend.

Phi language model executed in the browser using the Ratchet WebGPU runtime.

Pixel-space generation & general vision learner

Applies the look of a style image to your content image.

Draws and exports masks to use in inpainting workflows.

Web calculator for estimating and comparing LLM API costs across models.

Generates and edits images with the Qwen-Image-2.1 model.

Reads Ukrainian text aloud with a choice of voices.

CPU-based emotional multilingual text-to-speech with game-character style voices.

Pixel-diffusion decoder that upscales images.

Camera Control Dolly [Distilled]

Turns an uploaded image into a short animated clip.

Danbooru artist tag classification

Layout-controlled comics with a draggable bbox canvas

LTX-2.5 with Studio UI, scenes, LoRA tools and audio

Turns a text prompt into a short video with audio in four steps, using the MiniMax-H3 FlashGen LoRA.

All-in-one hub of general purpose tools useful for any agent

Strips the background from a photo.

Writes a GitHub issue draft with SmolLM2 running locally in the browser on WebGPU.

Builds a text corpus or dataset on a topic with an LLM and pushes it to your Hugging Face account.

Chat-style Llama assistant that helps turn a rough idea into a more detailed image-generation prompt.