MultiMatte
Open-source, promptable background-removal model — name the object to keep, it cuts everything else
Data verified Sep 30, 2026
Score
- Price
- Free
- Try it free
- Free plan
Editorial Verdict
AI-researched, not hands-on testedMultiMatte's prompt-steerable alpha matting is a genuine technical step up from typical one-size-fits-all background removers, and publishing both the weights and the inference library openly makes it one of the more developer-friendly image-AI releases in this category.
Reviewed by AIToolPundit Editorial Team · last verified September 2026
About MultiMatte
MultiMatte, released by Feyn Labs (Feyn Research), is an open-source background-removal model that can be aimed with a plain-English phrase: tell it "the dog" or "the jeans" and it keeps only that object, producing a smooth alpha matte instead of a hard-edged cutout. It's built on Meta's SAM 3 via low-rank (LoRA) fine-tuning that updates only 19.49M of SAM 3's 860M parameters (2.27% of the weights), yet substantially improves segmentation quality — scoring a 0.901 S-measure on the DIS-VD benchmark versus SAM 3's 0.667, with similar gains across eleven other benchmark splits. Unlike SAM 3's binary inside/outside masks, MultiMatte assigns each pixel a continuous opacity value, which is what lets it handle fine, translucent detail like hair or a blurred screen edge that boolean masks lose. The model weights are published on Hugging Face (feyninc/multimatte) and run through Feyn's own open-source nobg Python library (pip install nobg), with the LoRA adapter already merged into the released weights so no separate adapter library is needed at inference time. It can also be tried directly in the browser at usefeyn.com/multimatte with no installation.
Screenshots

Commonly Cited Strengths & Limitations
Strengths
- Free, open weights with a simple pip-installable library — no API key or account needed to self-host
- Meaningful, benchmarked improvement over the base SAM 3 model it's built on, not just a marketing claim
Limitations
- Best used by teams comfortable running a Python model locally rather than a point-and-click tool
Details
- Pricing Model
- Free
- Starting Price
- Free
- Team Size
- 2-10
- Category
- AI Image & Video Generation
Frequently Asked Questions
Is MultiMatte free?
Yes, MultiMatte is free to use.
Does MultiMatte have an API?
MultiMatte does not currently offer a public API.
Is MultiMatte open source?
Yes, MultiMatte is open source.
What are the downsides of MultiMatte?
Best used by teams comfortable running a Python model locally rather than a point-and-click tool
How does MultiMatte compare to Dreamlux?
MultiMatte and Dreamlux are both used for ai image & video generation. See the full side-by-side comparison for pricing and feature differences.
Key Features
- Promptable background removalNames the specific object to keep in plain English rather than removing all backgrounds indiscriminately.
- Alpha matting, not binary masksAssigns continuous opacity per pixel, handling fine detail like hair and translucent edges that hard masks lose.
- Built on SAM 3 via LoRAFine-tunes only 2.27% of Meta's SAM 3 weights, keeping its original text alignment while adding matting ability.
- Open weights + open libraryWeights are on Hugging Face and the inference library (nobg) is pip-installable and open source.
- No-install browser demoCan be tried directly at usefeyn.com/multimatte without any setup.
Technical Details
- Free Plan
- Available
Official Pricing
MultiMatte's published pricing · verified Sep 30, 2026
Free
Pricing shown above is sourced from MultiMatte's own published plans as of Sep 30, 2026 and can change at any time -- confirm current pricing on MultiMatte's own site before purchasing.
Real Pricing from Verified Users
See what people actually pay for MultiMatte
Reviews
0 reviews
No reviews yet. Be the first to share your experience!
Switching Stories
Real migration experiences with MultiMatte
Discussion
Ask questions, share tips, or discuss this tool with other users.
No discussion yet. Start the conversation!
Similar tools in AI Image & Video Generation
Other listed tools in AI Image & Video Generation, not reviewed as direct alternatives

Browser-based AI video generator that turns text or images into watermark-free videos in minutes.
Similar tool in AI Image & Video Generation
- Text to Video
- Image to Video
- Text to Image
AI GIF generator that turns written prompts into animated loops
Similar tool in AI Image & Video Generation
- Prompt-to-GIF generation
- Multi-format output
- Resolution and quality controls
Turns 12-20 selfies into 40+ studio-style AI portraits in under 10 minutes from $29
Similar tool in AI Image & Video Generation
- Custom Flux LoRA model per user
- 40+ portraits at 2048x2560 px
- 30+ suit styles and backgrounds
Turn a sentence, image, or link into realistic AI human-led product videos.
Similar tool in AI Image & Video Generation
- Text-to-video
- Image-to-video
- Link-to-video

AI platform for generating and retargeting 3D human motion for games, simulation and robotics
Similar tool in AI Image & Video Generation
- Video-to-motion
- Text-to-motion
- Locomotion

AI video editor that produces fully-edited videos from footage or prompts
Similar tool in AI Image & Video Generation
- AI Edit
- AI avatars and digital twin
- Auto captions and translation
People Also Viewed
More in AI Image & Video Generation
All-in-one audio & video editing

Open-source image generation AI
