Search the full model catalog.

Writes one detailed English paragraph describing a video clip from start to finish. THUDM built it to caption training data for its CogVideoX text-to-video model.

Meta's 7B Llama 2 chat model, aligned for English dialogue with supervised fine-tuning and RLHF. Text in, text out, with a 4k-token context window.

Cuts the main subject out of a photo and returns it on a transparent background at the original size. BiRefNet is a high-resolution segmentation model that uses gradient maps to keep fine edges.

Turns a text prompt into an image in 1 to 4 sampling steps. Black Forest Labs released it under Apache 2.0, so the model and its images can be used commercially.

Runs a vision language model of your choice on an image and returns its answer to your question as text. LLaVA v1.6 Mistral 7B is selected by default.

Hotshot-XL is a tool that allows you to generate GIFs from given text. Additionally, it can be controlled by another GIF.

Stability AI's 2-billion-parameter MMDiT text-to-image model from the SD3 series. It reads prompts through three text encoders, and Stability AI reports fewer spelling and spacing errors in rendered text.

It generates high-resolution images from text.

ConsistI2V animates a still image from a text prompt. Every generated frame attends to your image as frame one, and the starting noise borrows its low-frequency layout, so the clip stays close to the picture.

LLaVA is a tool that generates descriptions for a given image.

BLIP-2 paired with Meta's OPT-2.7B. Send an image and a question to get a short text answer or caption, returned with your question in front of it.

OpenAI's Whisper large-v3 turns an uploaded recording into a transcript. With diarization on, as it is by default, each segment gets start and end times and up to five speakers can be told apart.

Removes background noise from speech and field recordings with spectral gating, PCEN or a highpass filter, used alone or combined in one run.

Suno's Bark reads text aloud in 13 languages with one of 130 voice presets. Cues such as [laughs], [sighs] or lyrics wrapped in ♪ add nonverbal sounds and sung lines.

Voices a multi-speaker script with Suno's Bark. You start each line with a label such as sp_1 or sp_2 and pick a voice preset per label; the conversation comes back as one MP3.

Real-ESRGAN upscales real-world photos 2x or 4x and cleans up blur, noise and JPEG blocking without being told how the image was damaged. Faces can be restored with GFPGAN in the same run.