youcutai

AI models

Every model behind youcut.ai, with strengths, weaknesses and which tools use it.

removal

edit

Bria Eraser Bria AI via fal.ai

Bria Eraser is an inpainting model specialized for removal: given an image and a mask, it fills the region with background that matches surrounding texture, lighting and perspective rather than inventing new objects. It is the default engine for the Object Remover, People Remover, Text Remover and Watermark Remover on youcut.ai, with FLUX Fill as the fallback for very large regions.

FLUX.1 Fill [pro] Black Forest Labs via fal.ai

FLUX.1 Fill [pro] is Black Forest Labs' dedicated inpainting and outpainting model built on the 12B-parameter FLUX architecture. It can extend an image beyond its borders (outpainting) or regenerate a masked region with optional text guidance, keeping lighting, grain and perspective consistent with the original. On youcut.ai it powers the Image Expander, large-area removal fallback and scene generation in staging tools.

DDColor DAMO Academy (open source) via fal.ai

DDColor is an open-source image colorization model from Alibaba DAMO Academy that uses dual decoders, one for pixel-level color and one for semantic color queries, to produce vivid yet plausible colors. It performs especially well on historical photographs, assigning natural skin tones, skies and foliage. youcut.ai runs DDColor after CodeFormer restoration in the Colorize and Restore + Colorize modes.

CodeFormer S-Lab NTU (open source) via fal.ai

CodeFormer is a transformer-based face restoration model that uses a learned discrete codebook of facial features to reconstruct sharp, natural faces from degraded inputs. Its fidelity weight lets youcut.ai balance restoration strength against likeness, which is why restored relatives still look like themselves. It runs in Photo Restoration, as the face pass in the Photo Enhancer, and when face enhancement is enabled in the Image Upscaler.

Nano Banana (Gemini 2.5 Flash Image) Google via fal.ai

Nano Banana is the community name for Google's Gemini 2.5 Flash Image model. It became popular for keeping a person or product recognizable across edits, following multi-step natural-language instructions, and blending multiple reference images. It is fast (3–6 s) and reasonably priced. youcut.ai uses Nano Banana as the default engine for the AI Photo Editor, AI Background Generator, AI Headshot, Product Photo Generator and Product Staging, and offers its text-to-image mode in the AI Image Generator.

Kolors Virtual Try-On Kuaishou (Kolors) via fal.ai

Kolors Virtual Try-On is a diffusion-based garment transfer model from Kuaishou's Kolors team. It takes a person image and a garment image and renders the garment on the person with correct fit, drape and print placement, preserving the person's pose and identity. youcut.ai pairs it with a library of synthetic model presets so sellers can generate on-model images from flat lays without booking a shoot; Nano Banana is used for final harmonization.

upscale

image

FLUX.1 [schnell] Black Forest Labs via fal.ai

FLUX.1 [schnell] is the timestep-distilled variant of Black Forest Labs' 12B FLUX.1 model, producing high-quality images in as few as 1–4 diffusion steps. It is released under Apache 2.0 and costs around $0.003 per image to run, which lets youcut.ai offer it for a single credit. Quality is strong for its speed, with good prompt adherence and reasonable hands and text, making it the recommended starting point before switching to a premium model.

FLUX 1.1 [pro] Black Forest Labs via fal.ai

FLUX 1.1 [pro] is the high-end text-to-image model from Black Forest Labs, roughly six times faster than the original FLUX.1 pro with improved image quality, prompt adherence and diversity. It excels at photorealistic scenes, natural lighting, skin and materials, and supports up to 2K output in ultra mode. youcut.ai offers it for 3 credits per image and uses it as the photographic engine behind the AI Headshot tool.

Seedream 4.0 ByteDance via fal.ai

Seedream 4.0 is ByteDance's image model that combines text-to-image generation and instruction-based editing in one architecture, with native 2K resolution and multi-reference support. It is known for polished aesthetics, accurate bilingual (Chinese/English) text and strong consistency when editing. youcut.ai uses the text-to-image endpoint in the AI Image Generator and the edit endpoint in the AI Photo Editor, Image Expander and Product Photo Generator.

GPT Image OpenAI

GPT Image is OpenAI's natively multimodal image generation and editing model, the successor line to DALL·E. Its standout capabilities are precise instruction following, world knowledge, and rendering legible multi-line text, which makes it the best choice for posters, packaging mockups, infographics and complex edits described in several steps. It is the premium option on youcut.ai at 5 credits per image in both the AI Image Generator and the AI Photo Editor.

video

Wan 2.2 A14B Alibaba (open source) via fal.ai

Wan 2.2 A14B is Alibaba's open-source video generation model using a mixture-of-experts design (14B active parameters) that improves motion quality and cinematic control over Wan 2.1. It supports text-to-video and image-to-video at 480p and 720p for 5-second clips. Because it is open and efficient, youcut.ai offers it as the economy tier at 8 credits per clip, ideal for social content and product loops.

LTX-2 (13B distilled) Lightricks via fal.ai

LTX-2 is Lightricks' DiT-based video model; the 13B distilled variant is optimized for speed, producing 5–10 second clips in well under a minute, sometimes in seconds. It supports image-to-video with synchronized audio in some configurations and outputs up to 1080p. Quality is solid for its speed and price, which is why youcut.ai positions it as the fast/draft tier at 6 credits.

Kling 2.5 Turbo Pro Kuaishou via fal.ai

Kling 2.5 Turbo Pro is Kuaishou's flagship video model at the time of writing, combining realistic physics, expressive character motion and 1080p output with faster generation than previous Kling releases. It supports text-to-video and image-to-video at 5 or 10 seconds with reliable camera-move control. youcut.ai uses it as the primary "quality" tier for the AI Video Generator and Image to Video at 15 credits per clip, and its lip-sync sibling powers the Talking Photo tool.

Seedance 1.0 Pro ByteDance via fal.ai

Seedance 1.0 Pro is ByteDance's video generation model notable for native multi-shot output: a single prompt can produce a short sequence with consistent characters and coherent cuts. It offers strong prompt adherence, smooth motion and 1080p resolution for 5–10 second clips at a mid-range price. youcut.ai includes it in the AI Video Generator and Image to Video at 12 credits per clip.

Veo 3.1 Google DeepMind via fal.ai

Veo 3.1 is Google DeepMind's flagship video model. It produces 8-second 1080p clips with natively generated audio, including ambient sound, effects and lip-synced dialogue, and offers the most photorealistic output and best physical consistency currently available. It is expensive upstream, so youcut.ai reserves it for the premium tier at 60 credits per clip, available on paid plans.

lipsync