Everything in the catalog

Chat with a Llama tuned to bring the Eiffel Tower into almost every answer.

Extracts line-art sketches from an uploaded image and produces a descriptive prompt.

Answers questions about uploaded or YouTube audio, such as genre, key and instruments.

Official demo of Qwen-Image-2.1 for text-to-image creation.

Animates a still image into a 480p video clip.

Near real-time text-to-image and image editing with a 2-step Qwen-Image-2.1 LoRA.

Converts photos into anime-style artwork.

Text-to-speech (TTS) with Next-gen Kaldi

Talk to Qwen2.5-Omni with text, images, audio or video and get text or speech back.

Generates images that keep a consistent character, combining Kolors with Flux.

Text-to-speech demo for the Qwen3-TTS model family.

Open-weight text-to-speech that downloads the model once and runs in your browser.

Converts text to speech with the accelerated Kokoro-82M v1.0 model.

Generates images from text prompts, including text rendered inside the picture.

Writes a text caption describing the sound of a music clip.

Runs Microsoft's Florence-2 vision model in the browser for captioning, detection and OCR tasks.

Text-to-speech tool where a written description controls the voice, using the Parler-TTS model.

Translates text between many languages with an NLLB model executed inside the browser through Transformers.js.

Edits portrait photos with Kontext and selectable LoRA styles.

Checks whether a face in a photo or camera feed is a live person or a spoof, using passive 3D liveness detection.

Edits videos with the Wan 2.2 Animate model.

An interactive guide to 3D representations.

Calculates VRAM needs for a GGUF model, including splitting layers across GPU and CPU.

Makes scannable QR codes styled with AI art from a prompt and the text or URL to encode.