Clipdrop
Last reviewed
Apr 28, 2026
Sources
20 citations
Review status
Source-backed
Revision
v3 ยท 3,075 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Apr 28, 2026
Sources
20 citations
Review status
Source-backed
Revision
v3 ยท 3,075 words
Add missing citations, update stale details, or suggest a clearer explanation.
Clipdrop is an AI-powered image editing and creation platform developed by the French startup Init ML. Founded in 2020 by Cyril Diagne, Damien Henry, and Jonathan Blanchet, the platform grew out of a viral augmented reality prototype called AR Cut Paste that demonstrated the ability to capture real-world objects with a smartphone and paste them directly into desktop applications. Over the following years, Clipdrop evolved into a comprehensive suite of AI-driven visual tools, including background removal, object cleanup, image upscaling, relighting, text removal, uncropping, and text-to-image generation powered by Stable Diffusion. The platform accumulated over 15 million users before being acquired by Stability AI in March 2023 and subsequently sold to Jasper AI in February 2024.
Clipdrop traces its origins to the work of Cyril Diagne, a French creative coder and former Artist in Residence at the Google Arts & Culture Lab (2015 to 2020). Diagne studied at Gobelins, l'ecole de l'image in Paris and co-founded the interactive art collective LAB212 after graduating in 2008. He later served as Professor and Head of Media & Interaction Design at ECAL in Lausanne, Switzerland.
In May 2020, Diagne published a prototype called AR Cut Paste on GitHub. The tool used a combination of BASNet (Boundary-Aware Segmentation Network), a deep learning model for salient object detection, and OpenCV SIFT (Scale-Invariant Feature Transform) to enable users to point a smartphone camera at a real-world object, automatically segment it from the background, and paste it into a desktop application such as Adobe Photoshop. The prototype took approximately 2.5 seconds to cut an object and about 4 seconds to transfer it to the desktop. The demo went viral on Twitter, attracting attention from Adobe's official Photoshop account and generating widespread media coverage.
The success of AR Cut Paste motivated Diagne to formalize the concept into a product. In July 2020, he co-founded Init ML alongside Damien Henry and Jonathan Blanchet, both of whom had backgrounds at Google. Henry had worked in operations and product roles, while Blanchet brought expertise in engineering and machine learning. The company was headquartered in Paris, France.
The founding vision was to simplify a workflow that traditionally required multiple steps: photographing an object, transferring the file to a computer, opening an image editor, and manually removing the background. Clipdrop compressed this entire process into a single "point, capture, drop" interaction.
Init ML was accepted into Y Combinator's Winter 2021 (W21) batch, giving the startup access to Silicon Valley mentorship, network connections, and global visibility. During its time in the accelerator program, the team refined Clipdrop's core product and expanded its feature set beyond the original AR copy-paste functionality.
In August 2021, Init ML closed a seed funding round of approximately 1.2 million euros (roughly $1.5 million) led by Air Street Capital, a venture firm focused on artificial intelligence. The round included participation from 11 investors, among them Kima Ventures and SOMA. Several prominent technology executives participated as angel investors:
| Investor | Role / Background |
|---|---|
| Maria Raga | Former CEO of Depop (acquired by Etsy) |
| Kaj Drobin | Co-founder of TicTail (now VP at Shopify) |
| Siavash Ghorbani | Co-founder of TicTail |
| Carl Riviera | Co-founder of TicTail |
| Edward Miller | CEO of Scape (acquired by Facebook) |
| Namrata Ganatra | Former CTO of Lambda School |
| Ian Hogarth | CEO of SongKick |
| Matt Robinson | CEO of Nested |
The seed capital supported the expansion of Clipdrop's product line and its growing user base. By mid-2021, users had already created over 5 million "clips" (object captures and edits) within six months of the platform's initial launch.
Clipdrop launched initially as a focused AR-based object capture tool but rapidly expanded into a full suite of AI image editing tools. By 2023, the platform offered nine primary tools, each powered by distinct AI models.
The background removal tool was Clipdrop's flagship feature from the start. It uses image segmentation models to identify and isolate the foreground subject from the rest of the image. The tool supports a wide range of subjects including portraits, products, and complex scenes. Clipdrop marketed it as "the most accurate background removal solution available on the market," though independent benchmarks from Photoroom and others showed that all background removal tools (including Clipdrop, Remove.bg, and Meta's Segment Anything Model) struggled with particularly fine details such as individual hair strands, bicycle spokes, and lace textures.
The Cleanup tool, also available at the standalone domain cleanup.pictures, allows users to remove unwanted objects, people, text, and defects from images. Unlike traditional clone stamp or content-aware fill tools found in conventional image editors, Clipdrop's Cleanup feature uses AI to intelligently predict what should appear behind the removed element. Users simply brush over the area they want removed, and the AI fills the region with contextually appropriate content. Pro subscribers gained access to an HD mode for higher-resolution output.
Clipdrop's upscaling tool enhances image resolution by 2x or 4x while reducing noise and recovering fine details. The technology uses super-resolution algorithms trained on large datasets of image pairs to predict high-frequency details that are absent in lower-resolution originals. Pro plan subscribers could access upscaling up to 16x the original resolution.
Relight is a tool that allows users to adjust the lighting in a photograph after it has been taken. The tool uses AI-based depth estimation to understand the three-dimensional structure of the scene, then applies virtual light sources that cast realistic shadows and highlights. Users can position light sources interactively within the interface, and the AI recalculates the shading accordingly. The feature proved popular with product photographers, portrait editors, and digital artists who needed to adjust lighting conditions without re-shooting.
The Text Remover tool detects and removes text from images. It uses optical character recognition (OCR) and inpainting models to first identify text regions and then fill those regions with content that matches the surrounding area. Common use cases include removing watermarks, cleaning up photographs of signage, and preparing images for localization where text needs to be replaced in different languages.
Uncrop extends the boundaries of an image by generating new content beyond the original frame. Built on foundation models from Stability AI, the tool predicts what should exist outside the visible edges of a photograph. Users upload an image and select a target aspect ratio, and the AI generates the missing portions. Practical applications include converting square Instagram photos to widescreen formats, restoring cropped areas in old photographs, and adapting images for social media platforms that require specific dimensions.
Replace Background combines Clipdrop's background removal capability with generative AI. After isolating the foreground subject, the tool generates a new background based on a text prompt or selected scene description. This feature was released in beta and enabled users to place products or people into entirely new environments without manual compositing.
Clipdrop's text-to-image tool generates images from written prompts. The feature is built on top of Stable Diffusion and was optimized for speed within the Clipdrop interface. Users can choose from multiple artistic styles, including photographic, anime, digital art, comic book, fantasy art, analog film, neon punk, isometric, low poly, origami, line art, cinematic, and 3D model. Available aspect ratios range from ultrawide (21:9) to vertical (9:16). Each prompt generates four image variations for the user to choose from.
The Universal Resizer adapts images automatically for social media formats. Rather than simply cropping or stretching, the tool intelligently adjusts composition and fills in missing areas to fit the target dimensions for platforms such as Instagram, Facebook, Twitter, LinkedIn, and YouTube.
After Stability AI acquired Clipdrop in 2023, the platform became a primary consumer-facing showcase for Stability AI's generative models.
Clipdrop served as the official demo platform for Stable Diffusion XL (SDXL), released in July 2023. SDXL represented a major upgrade over previous Stable Diffusion versions, generating images at 1024x1024 resolution with a two-stage pipeline (base model and optional refiner). Through Clipdrop, users gained free access to SDXL without needing to install software or manage GPU resources locally.
In late May 2023, Stability AI launched Reimagine XL on Clipdrop. This tool generates multiple variations of an uploaded image, preserving the overall composition and subject matter while changing specific details, colors, and artistic elements. The technology is based on an open-source Stable unCLIP model that replaces Stable Diffusion's text encoder with an image encoder. Instead of conditioning generation on a text prompt, Reimagine XL encodes the source image into a latent representation, adds controlled noise, and then generates new images that are visually similar but not pixel-identical to the original. This approach differs from standard image-to-image processing because the source image is fully encoded first, meaning the generated output does not reuse any pixels from the original.
The initial version, Stable Diffusion Reimagine, launched in March 2023. The XL upgrade in May 2023 brought higher-resolution output and improved quality.
In November 2023, Clipdrop became the demo platform for SDXL Turbo, a distilled version of SDXL 1.0 trained using Adversarial Diffusion Distillation (ADD). SDXL Turbo could generate images in a single network evaluation step, enabling near-real-time text-to-image synthesis. Users could type a prompt and see results appear almost instantly, a significant departure from the multi-second generation times of standard diffusion models.
Clipdrop was designed as a cross-platform product from its earliest days. The platform was available through multiple channels:
| Platform | Description |
|---|---|
| Web application | Full-featured browser-based access at clipdrop.co |
| iOS app | Mobile application for iPhone and iPad |
| Android app | Mobile application for Android devices |
| macOS application | Native desktop app for Mac |
| Windows application | Native desktop app for Windows |
| Photoshop plugin | Direct integration with Adobe Photoshop, allowing users to drop captured objects as new layers with editable masks |
| Figma plugin | Integration with Figma for background removal, object removal, text removal, and automatic image decomposition into layers |
The Figma plugin was free to use with unlimited image processing and required no subscription. The Photoshop plugin allowed users to insert captured objects directly as new layers with editable masks, fitting naturally into existing design workflows.
Clipdrop also supported integration with other applications including Microsoft PowerPoint, Google Docs, Canva, and Pitch.
Clipdrop offered a REST API that provided programmatic access to its core image processing capabilities. The API endpoints included:
| API Endpoint | Function |
|---|---|
| Remove Background | Segment and extract foreground subjects |
| Cleanup | Remove unwanted objects, text, or people |
| Image Upscaling | Enhance resolution up to 16x |
| Text to Image | Generate images from text prompts using Stable Diffusion |
| Remove Text | Detect and remove text from images |
| Uncrop | Extend image boundaries with AI-generated content |
| Replace Background | Combine background removal with generative fill |
The API was used by developers and businesses to integrate Clipdrop's capabilities into their own applications, e-commerce platforms, and content management systems. After the Jasper acquisition, the API was migrated to the Jasper API platform.
Clipdrop operated on a freemium model with three tiers:
| Plan | Price | Features |
|---|---|---|
| Free | $0 | Limited daily usage, watermarked results, standard resolution, 100 images per month |
| Pro | $9/month ($7/month billed annually) | Unlimited access, HD resolution, no watermarks, batch processing (up to 10 images), 16x upscaling, Relight templates, priority processing |
| API / Business | $29/month | API access, 5,000 image allowance, developer documentation, commercial integration support |
All plans permitted commercial use of generated and edited images.
Clipdrop earned significant recognition in the product and technology community:
On March 7, 2023, Stability AI announced the acquisition of Init ML, the company behind Clipdrop. The financial terms of the deal were not disclosed. At the time of the acquisition, Clipdrop had amassed more than 15 million users.
Init ML became a wholly owned independent subsidiary of Stability AI, with all employees retained. The three co-founders took on new roles within the Stability AI organization:
| Co-founder | Previous Role | New Role at Stability AI |
|---|---|---|
| Cyril Diagne | CEO, Init ML | Vice President of Innovation, Clipdrop |
| Damien Henry | COO, Init ML | Senior Vice President of Product, Clipdrop |
| Jonathan Blanchet | CTO, Init ML | Vice President of Engineering, Clipdrop |
The acquisition gave Clipdrop access to Stability AI's generative AI models, including Stable Diffusion XL and its successors. For Stability AI, Clipdrop provided a polished consumer-facing platform to showcase its models to a broad audience. During the Stability AI ownership period, Clipdrop launched several new features built on Stability AI technology, including Reimagine XL, SDXL-powered text-to-image generation, and SDXL Turbo real-time synthesis.
On February 20, 2024, Jasper AI, an AI marketing platform, announced it had completed the acquisition of Clipdrop from Stability AI. The deal closed less than a year after Stability AI's original purchase of Init ML. Financial terms were again not disclosed.
The acquisition was part of Jasper's strategy to build a comprehensive "marketing copilot" that combined text generation with image creation and editing capabilities. Jasper CEO Timothy Young stated: "Marketing is visual. The addition of Clipdrop to Jasper will advance our vision to be the most comprehensive end-to-end marketing copilot in the industry, powering all the formats, channels, and functions enterprise marketing teams need."
Emad Mostaque, then CEO of Stability AI, expressed support for the deal, noting that Jasper had been "a valued partner, leveraging Stability AI models for many years" and that Stability AI would "continue to partner with Clipdrop on research and deliver our cutting-edge models to their platform."
Clipdrop co-founder Damien Henry took on an expanded role at Jasper, leading research and innovation across the entire Jasper product line. The Clipdrop platform continued to operate as a standalone product at clipdrop.co, while its API was integrated into the Jasper API for business customers.
Clipdrop's tools rely on several categories of AI models:
The background removal and object isolation features use image segmentation models that classify each pixel in an image as either foreground or background. The original AR Cut Paste prototype used BASNet (Boundary-Aware Segmentation Network), a predict-refine architecture that combines a densely supervised encoder-decoder network with a residual refinement module. Later versions of Clipdrop likely incorporated more advanced segmentation architectures as the field progressed.
The Cleanup and Text Remover tools use inpainting models that fill in masked regions of an image with content that is contextually consistent with the surrounding pixels. These models are trained on large datasets of images with artificially masked regions, learning to reconstruct the missing portions in a way that appears natural.
The Image Upscaler uses super-resolution neural networks that learn mappings between low-resolution and high-resolution image pairs. These networks predict high-frequency details (textures, edges, fine patterns) that cannot be recovered through simple interpolation methods like bicubic scaling.
The Relight tool uses monocular depth estimation models to infer the three-dimensional structure of a scene from a single two-dimensional image. This depth map enables the placement of virtual light sources that produce physically plausible shadows and highlights.
The text-to-image, Reimagine, Uncrop, and Replace Background features all rely on diffusion models, specifically variants of Stable Diffusion. These models work by learning to reverse a gradual noising process: during training, they learn to remove noise from progressively noisier versions of images, and during inference, they start from pure noise and iteratively denoise it into a coherent image guided by a conditioning signal (text prompt, image encoding, or partial image).
As of early 2026, Clipdrop continues to operate at clipdrop.co under Jasper's ownership. The platform maintains its suite of AI image editing tools, with pricing starting at a free tier (100 images per month) and paid plans beginning at $9 per month. The Clipdrop API has been incorporated into Jasper's broader API platform, though Stability AI has continued to provide its generative models to the Clipdrop platform as part of an ongoing partnership.
The competitive landscape for AI image editing has intensified since Clipdrop's founding, with tools such as Adobe Firefly, Canva AI, Photoroom, and Remove.bg offering overlapping capabilities. The broader trend in the market has moved toward integrating AI editing tools directly into existing creative software platforms rather than offering them as standalone web applications.