From FaceU to CapCut to AI Video: The Team Behind ByteDance’s Imaging Apps Bets on Flova

The team behind FaceU and products that helped shape ByteDance’s consumer imaging ecosystem is entering a new chapter in video creation.

Founded by Guo Lie, creator of FaceU and former leader of ByteDance’s imaging business, Flova.ai has raised more than $80 million across two funding rounds, with backing from Sequoia Capital, IDG Capital and Yunji Capital. The company launched globally in October 2025 and is now building what it describes as an AI-native approach to video production.

Before Flova, Guo founded FaceU, which was acquired by ByteDance for approximately $300 million in 2018.

He later led ByteDance’s imaging business and was involved in the incubation and development of CapCut.

With Flova, the team is turning its attention from consumer imaging to a broader question: what happens when AI stops being a generation tool and becomes part of the production team?

AI Video Has Solved Generation. Production Remains the Problem.

The past few years have seen rapid advances in AI video generation. Platforms including Runway, Pika, Kling and Sora have made it possible to generate increasingly sophisticated images and video from natural-language instructions.

But generating individual clips is only one part of making a video.

A real production requires a creator to manage scripts, characters, shots, references, assets, model selection, versions, revisions and editing—often across multiple tools.

Flova takes a different approach.

Available at Flova.ai, the web-based platform brings multiple leading image and video models into a single creative environment, while placing an AI Agent at the center of the production workflow.

Rather than treating each generation as an isolated request, Flova is designed to understand the relationships between a project's script, shots, assets and timeline.

That distinction becomes increasingly important as projects grow more complex.

An Agent That Understands the Project, Not Just the Prompt

A prompt describes a request. A project contains context.

Characters have identities and relationships. Stories have continuity. Brands have visual rules. Assets have different versions. A change to one reference may affect multiple shots.

Flova's Agent is designed to retain and work with this project-level context.

Creators can upload long-form scripts, character profiles, world-building materials, brand guidelines and production requirements. Flova currently supports up to 100,000 Chinese characters in a single script upload, allowing creators to provide the Agent with an entire story while controlling which episode, scene or shot they want to produce.

The Agent can then help transform that context into editable storyboards, organize and bind assets, prepare generation prompts, coordinate AI models, manage revisions and assemble rough timelines.

The objective is not simply to build an AI assistant that can answer questions about a video.

It is to create an Agent that can work with the structure of the project itself.

From Prompt Templates to Creative Skills

Flova's latest 1.0 release introduces another layer: Flova Skills.

Professional creators can create Skills that capture their preferred workflows, prompting structures, visual standards, creative preferences and production methods.

A Skill is designed to be more than a reusable prompt. It can encode a repeatable way of working that an Agent can apply to future projects.

A filmmaker could build a Skill around a particular cinematic workflow. A commercial creator could capture a brand's visual production standards. An AI creator could turn a proven prompting and iteration process into a reusable creative system.

Flova currently offers a video-focused Skill Hub with more than 100 professional Skills, while also building a community where experienced creators can share their methods with others.

This creates a different relationship between expertise and AI:

Models provide generation capabilities.

 Skills capture creative methodology.

 Agents execute the workflow.

 Creators remain responsible for creative judgment.

A Multi-Model Production Environment

Flova is also designed around a multi-model workflow rather than tying creators to a single generation model.

Creators can access leading AI image and video models from within the same production environment, while Flova manages the surrounding project structure.

This means creators can focus less on moving assets and prompts between different AI products and more on deciding which creative direction works.

The company sees this as an important distinction between an AI video generator and an AI video production platform.

The former answers:

“Can AI generate this shot?”

The latter needs to answer:

“How does this shot fit into the project, what assets should it use, what happens when it changes, and how does the project move forward?”

Flova is built around the second question.

Making Professional Workflows More Accessible

Flova's ambition is not to replace creative judgment.

Creators still decide what the project should look like, which direction is right and which result is worth keeping.

The Agent handles more of the operational work surrounding those decisions—from organizing context and preparing prompts to coordinating generations, managing assets, tracking versions and supporting revisions.

For professional creators, this can reduce repetitive production work.

For less experienced creators, Skills can provide access to workflows and methods that would otherwise take years to develop.

The result is a different model of AI-assisted creation:

The creator brings the vision. The Agent helps carry it through production.

Building a Creative System That Learns Over Time

Flova's longer-term vision extends beyond a single generation session.

Approved characters, products, environments and references can be retained for future projects. Creative workflows can be turned into Skills. Project standards can be updated and reused. Previous versions can remain available rather than being overwritten.

Over time, the production system becomes more valuable because it accumulates the creator's assets, methods and decisions.

This is the foundation of Flova's approach to Agent-Native Video Production: moving AI video from a sequence of disconnected generations toward a continuous production environment where context, creative methods and project knowledge can be reused.
 

The Next Chapter for AI Video

The first wave of generative AI made it possible to create individual images and clips with increasingly simple instructions.

Flova is betting that the next wave will focus on something broader: making the production process itself AI-native.

That means moving beyond asking AI to generate a shot and toward giving an Agent enough context, tools and creative methodology to help move an entire project forward.

For the team that previously helped bring FaceU and ByteDance's imaging products to hundreds of millions of users, Flova represents a new chapter in the same long-running question:

How can technology make sophisticated visual creation accessible to more people?

This time, the answer may not be another camera, editor or generation model.

It may be an Agent that works alongside the creator.

About Flova

Flova.ai is an AI-native video creation platform focused on Agent-driven production workflows. Launched globally in October 2025, Flova combines leading AI image and video models with project memory, contextual understanding, intelligent asset management, timeline workflows and reusable Skills.

Flova was founded by Guo Lie, creator of FaceU and former leader of ByteDance's imaging business. FaceU was acquired by ByteDance for approximately $300 million in 2018. Guo was subsequently involved in the incubation of products including CapCut, Hypic and BeautyCam.

Flova has raised more than $80 million across two funding rounds, backed by Sequoia Capital, IDG Capital and Sky9 Capital.

Email: contact@flova.ai

Website: www.flova.ai