Reve ships Chrome extension to pull web images into AI projects
Updated July 27th, the browser tool saves web images as references, palettes, templates and album items inside Reve.
By Ryan Merket · Published
Why it matters
The extension turns the open web into an intake layer for Reve, expanding the product from AI image generation into the daily research and organization work that precedes creation.
Reve announced its Chrome extension on Wednesday, giving users a browser-level tool for saving images from across the web directly into its AI creation and editing platform.
https://x.com/reve/status/2082513253457158421?s=46
The Palo Alto developer was founded by Christian Cantrell, a former Stability AI product chief who spent more than 17 years at Adobe, alongside researchers Michael Gharbi and Taesung Park, technology executive Jon Watte and investor Mike Speiser. Cantrell says he helped launch Stability AI's platform, API, DreamStudio 1.0 and integrations before starting Reve in March 2023.
Turning the browser into an input layer
The Chrome Web Store listing says the extension places a Reve icon over images when a user hovers over them. Users can then save an image as a creative reference, extract a color palette, add it to an album or convert it into a template for generating variations.
Those items sync into a user's existing Reve projects. The extension requires a Reve account and can access website content, according to the listing. Version 1.0.13 was updated on July 27th and had 32 users and two ratings as of July 29th, indicating that distribution remains at an early stage.
The extension expands Reve's product beyond the generation prompt and editing canvas. Browsing for visual references is already part of many design workflows. Reve is placing itself at the point where that research becomes structured input for an AI project, reducing the steps between finding an image and using its composition, colors or style as guidance.
That intake layer fits Reve's broader product thesis. Reve has built its image platform around a structured layout representation that breaks generated pictures into addressable objects, text and regions. Users can move, resize, replace or rewrite parts of a composition without starting another generation from scratch.
When Reve released Reve 2.0 on June 3rd, Reve said the model generated native 4K images and used its layout system to support direct manipulation. Reve also claimed the model debuted at No. 2 on Arena's text-to-image leaderboard and was trained with 10 times fewer GPUs than competing systems. Those performance and training claims were supplied by Reve.
Gharbi previously worked as a research scientist at Adobe Research after completing a PhD at MIT CSAIL. Park also worked at Adobe Research and earned his computer science PhD at the University of California, Berkeley. Their research backgrounds help explain Reve's focus on the model and the editor as one product, rather than treating generation as a separate service beneath a conventional design interface.
A product launch after the OpenAI deal
The extension arrived two days after Reve announced an investment from OpenAI. Reve said members of its AI research group would join OpenAI to work on multimodal research, while Reve would remain independent and continue building with models from multiple providers.
The arrangement gives Reve additional resources while transferring part of its research talent to one of the largest competitors in AI image generation. Reve's next test is therefore increasingly a product and distribution challenge: converting its remaining technical work into a creative system that users return to for collecting references, generating assets and making precise edits.
Seed investor Basis Set Ventures describes Reve as a multimodal image creation and editing platform built around creative intent and direct manipulation. The investor lists its backing as a 2024 seed investment. Reve has not published financial terms for the OpenAI investment.
The Chrome extension also puts image rights closer to the point of import. Saving a picture from a webpage does not confer permission to reproduce or adapt it. Reve's usage policy prohibits infringement of third-party intellectual property, leaving creators responsible for how imported references, templates and generated variations are used.
For Reve, the browser tool is a compact distribution move with a larger purpose. Each saved reference gives users another reason to organize visual research inside Reve before generation begins. That can make Albums, References and Templates part of the user's continuing workflow, rather than features encountered only after opening the image generator.