AI Model Catalogue

GeoCalib (ECCV 2024): Single-image camera calibration. Estimates focal length, FoV, distortion, roll and pitch from one image using a deep net + Levenberg-Marquardt optimizer. Works on both outdoor and indoor scenes.

Super video quality enhancement featuring fast upscaling with TensorRT and frame interpolation with RIFE.

A custom Flux LoRA model trained on painterly illustrated poster art inspired by Blade Runner 2049. The style features atmospheric cyberpunk cityscapes with dramatic scale — tiny silhouetted figures dwarfed by massive holographic projections and towering

New and improved version of Veo 3 Fast, with higher-fidelity video, context-aware audio and last frame support

A sub 1 second 0.01$ multi-image editing model built for production use cases. For image generation, check out p-image here: https://replicate.com/prunaai/p-image

Generate professional e-commerce product photos from a single image. Automatically removes background, creates realistic studio scenes, and adds natural shadows.

The highest fidelity image model from Black Forest Labs

This is a version of the MoE Qwen 3.5 35B optimised by Pruna AI.

Create a dotted waveform video from an audio file

4x face image upscaler trained on FFHQ dataset using DAT (Dual Aggregation Transformer) architecture. Optimized for portrait and face photos.

This LUT-based color filter is ideal for color grading user-generated AI videos or short videos shot on smartphones.

RobustVideoMatting on Replicate: input mp4 video, output black-and-white alpha-mask.mp4.

Kling 2.5 Turbo Pro: Unlock pro-level text-to-video and image-to-video creation with smooth motion, cinematic depth, and remarkable prompt adherence.

Generate and edit images with Alibaba's Wan 2.7

Metric3D v2 (TPAMI 2024): Monocular metric depth and surface normals from a single image. Predicts real-world depth in meters. Works indoor and outdoor.

Max-quality image generation and editing with support for ten reference images

ML model that detects acne, dark circles, wrinkles and oily skin in one go.

Fast video generation with text-to-video, image-to-video, and start-end-to-video modes. Up to 16 seconds at 1080p with synchronized audio.

Generate videos from reference images or clips while preserving subject identity using Alibaba's Wan 2.7 reference-to-video model

Ultralytics YOLOv8s worldv2 Real-Time Open-Vocabulary Object Detection model with 12.7M parameters. Achieves 37.7 mAP50-95 on COCO dataset. Optimized for real-time inference

Monocular metric depth estimation for panoramic images

Seedream 4.5: Upgraded Bytedance image model with stronger spatial understanding and world knowledge

Beta/RFC version of https://replicate.com/nelsonjchen/op-replay-clipper

A visual style focused on modern enterprise architecture—reflective blue glass skyscrapers captured from upward perspectives, emphasizing symmetry, scale, and a clean, premium corporate aesthetic.

Reimagine any song in a different style — change voice, instruments, genre, and arrangement while keeping the original melody

Enables precise control of character actions and expressions from a reference image.

Image to separate stems from a song, using demucs and spleeter

Generate full-length songs up to 3 minutes from text prompts or images with Lyria 3 Pro, Google's most capable music generation model

Edit and transform videos with text prompts and reference images. Style transfers, object replacement, character transformation, and more.

New and improved version of Veo 3, with higher-fidelity video, context-aware audio, reference image and last frame support

This is an optimised version of the hidream-l1 model using the pruna ai optimisation toolkit!