Draft:Bria Artificial Intelligence

• Comment: In accordance with Wikipedia's Conflict of interest guideline, I disclose that I have a conflict of interest regarding the subject of this article. Ronimatza (talk) 22:52, 24 February 2026 (UTC)

• Comment: The references are in a bad way in terms of layout and parameters, such that it's difficult to ascertain notability and unfair on volunteer editors to have to sort the citations. See WP:REFB and kindly tidy up these citations. ChrysGalley (talk) 19:14, 5 February 2026 (UTC)

• Comment: Needs more significant coverage in particular. Brief mentions in minor awards isn't enough. Lynch44 18:08, 8 January 2026 (UTC)

---

---

Bria Artificial Intelligence

Bria (also stylized as BRIA or Bria AI) is an Israeli- American artificial intelligence company that specializes in visual generative AI technology for enterprise use. The company develops AI platforms for visual generation, editing, and content creation, with a focus on controllability, and directability of the visual. The company is known for a unique attribution engine that promises fully licensed data and training sets, and compensates data owners for the use of the data. Bria is headquartered in New York, with offices in London, Los Angeles and Tel Aviv.

Bria stands for the Hebrew word translation of Generation.

History

Bria was founded in 2020 by Dr. Yair Adato and Gal Jacobi. Dr. Adato serves as the company's CEO today, holds a PhD in Computer Vision and has been a vocal advocate for responsible AI development and enterprise IP compliance.

The company emerged with the goal of building a visual generative AI platform for commercial enterprise use, differentiating itself from the outset by training its models on high controllability and exclusively on licensed data rather than scraped web content. In October 2022, Bria announced a strategic partnership with Getty Images, one of the world's largest visual content providers, aiming to integrate responsible AI tools into Getty's platform. In 2024, Bria announced a strategic partnership with Publicis Groupe (which is one of its investors as well).

In November 2025, Bria launched Fibo, its flagship foundation model, introducing a new paradigm of structured, deterministic image generation powered by the company's proprietary Visual GenAI Language (VGL).

Platform and Products

Bria operates a Visual Platform-as-a-Service (Visual PaaS) serving developers, product teams, creative professionals, and AI agents as first-class users. The platform is organized around four product lines.

The Visual Engine is Bria's core offering - an API-first suite of over 30 image and video generation and editing capabilities, comparable in scope to a cloud-native Photoshop for programmatic use. It includes text-to-image generation, image editing (background removal, object addition/removal/replacement, relighting, restyling, inpainting, outpainting, resolution increase up to 8K), video editing, product shot APIs and tailored generation for training custom models on a customer's own visual IP. The Visual Engine is designed for embedding directly into products, workflows, and agent pipelines.

Ads & Catalog automate the creation of advertising and catalog assets at scale - generating localized, personalized visuals while preserving brand templates and product identity.

Brand Checker validates campaign assets for brand compliance, safety, and legal requirements through an API-first interface.

The Media Suite provides integrations with professional creative tools including Adobe Photoshop, Figma, Nuke, Houdini, Toon Boom, and ComfyUI, enabling AI-assisted generation and editing within established production pipelines.

Integration and Deployment

Bria is available through multiple integration paths: REST APIs, SDKs, an MCP server for AI agents, embeddable iFrames, ComfyUI nodes, and DCC plugins. The platform is natively listed on AWS and Azure marketplaces, with GCP also supported. For organizations with data sovereignty or regulatory requirements, Bria offers on-premises and Bring Your Own Cloud (BYOC) deployment. Enterprise customers can also access source code and model weights directly. All foundation models are openly available on Hugging Face for non-commercial use, with commercial licensing for production.

Visual GenAI Language (VGL)

Visual GenAI Language (VGL) is Bria's proprietary visual language specification for visual tasks. Rather than relying on free-text prompts, VGL encodes visual intent as explicit, machine-readable schemas - describing composition, objects, people, camera parameters, lighting, colors, sentiment, and style as independently controllable attributes. VGL is what enables Bria to deliver deterministic, reproducible, and auditable outputs - transforming visual AI from probabilistic approximation into a structured engineering discipline.

The Fibo Model Family

The Fibo model family comprise of multiple visual generative AI models sharing a common VGL control language, differing only in their execution characteristics. Some of the models in the family include: Fibo Generation, Fibo Edit, Fibo Brand, and Fibo Lite.

Licensed AI and Attribution

Bria's models are trained exclusively on licensed data sourced from over 30 content partners.Partners include major content providers such as Getty Images, Envato, Alamy, Freepik, and Depositphotos, among others.The company developed a patented attribution engine that tracks the contribution of individual training images to each generated output and programmatically compensates data owners based on their content's influence- a model Bria has described as analogous to music streaming royalty systems. As of 2025, Bria has filed 23 patents.

Michael Feinstein, Bria's CTO described the company's approach in an interview as making a deliberate choice not to adapt pre-existing foundation models trained on unlicensed data, instead training all models from scratch to ensure complete data provenance.

The company also claims to be the first AI provider to achieve full compliance with the EU AI Act.

Funding and Investors

Bria has raised a total of approximately $65 million across multiple funding rounds.

Early investors included Entrée Capital and IN Ventures (a venture arm of Sumitomo Corporation, Japan), who backed the company from its early stages.

Series A- (February 2024) The Series A was jointly led by GFT Ventures, Intel Capital, and Entrée Capital, with additional participation from Publicis Groupe, Getty Images, Samsung, IN Venture , Atinum Investment (South Korea), Z Venture Capital / LY Corporation (Japan), Mirae Asset Venture Investment, and J-Ventures.

Series B- (March 2025) The Series B was led by Red Dot Capital, with participation from Maor Investment, and existing investors.

HPA Awards and Recognition

In 2026, Bria AI received two Hollywood Professional Association (HPA) Awards at the 2026 HPA Tech Retreat in Rancho Mirage, California. The company won the HPA Award for Innovation in Pre-Production for its GenAI Attribution Technology, which traces training data influencing each AI-generated output and automatically compensates contributors based on their impact. Bria also received the HPA Transformative Impact Award, which recognizes technology with the highest potential to redefine the professional media content industry.

Notable Partnerships and Accelerator Programs

Warner Bros. Discovery -Innovate On The Lot Accelerator (2025)

Bria was selected as one of 12 global startups for Warner Bros. Discovery's 2025 "Innovate On The Lot" accelerator program, powered by Amazon Web Services and ArentFox Schiff.The program offered selected companies the opportunity to work directly with WBD business leaders toward building strategic partnerships with the global media and entertainment company.Bria's selection was reported by Israeli tech media as a significant recognition of the company's enterprise AI capabilities.

Deloitte Launchpad

Bria AI was selected as one of ten Israeli startups to participate in Deloitte Launchpad's 9th cohort, a U.S. go-to-market acceleration program designed to help Israeli growth-stage companies scale into the American market.

Intel Ignite

Bria AI was selected to participate in Intel Ignite's Fall 2020 cohort, an accelerator program for deep tech startups leveraging Intel's resources, mentorship, and global network.

Notable Use Cases

Beetlejuice Beetlejuice - Trailer Park Group / Warner Bros. Pictures (2024)

In 2024, entertainment marketing agency Trailer Park Group partnered with Bria to build the "Welcome to the Afterlife Generator", a user-generated content campaign promoting Warner Bros. Pictures' Beetlejuice Beetlejuice.The campaign allowed fans to upload selfies and receive personalized AI-generated portraits styled after characters from the film.

"Lidlize"- Lidl France & Marcel Worldwide (2024–2025)

The "Lidlize" campaign, developed by Paris-based creative agency Marcel (a Publicis Groupe subsidiary) for Lidl France, allowed users to type the name of any object and instantly see it transformed into Lidl's signature red, blue, and yellow brand palette. Lidl then manufactured the images generated into physical objects that were sold in their stores.

The campaign won:

• D&AD Wood Pencil
• Three Clio Awards
• Four Cannes Lions

Industry Recognitions and Awards

• CB Insights AI 100 (April 2025) -Named to CB Insights' ninth annual list of the 100 most promising private AI companies globally.
• TIME 100 Most Influential Companies (2026 nomination)- Nominated for TIME magazine's list of the 100 most influential companies.
• Forbes Israel Next Billion Dollar Startup (2025)- Selected for Forbes Israel's list identifying the most promising high-growth startups.
• Ynet Disrupt 40 -Dr. Yair Adato was named among the 40 most influential people driving innovation and vision in the Israeli tech industry.
• N12 / Mako Most Promising Startups of 2025- Named among the top promising Israeli startups of 2025 by Mako/N12.
• AWS Design Partner of the Year (2024)- Recognized by Amazon Web Services for work in visual generative AI solutions.

External Links

Bria Official Website

Bria’s models on Hugging Face

Bria’s models on Fal.Ai