Mosaic is a web app that uses AI agents to edit videos by executing tasks arranged in nodes.



Mosaic is a video editing platform that utilizes an AI agent. Unlike traditional timeline-based editing software, it employs an autonomous editing paradigm in which the AI agent autonomously performs editing tasks based on the user's instructions. It was developed by former Tesla engineers Adish Jain and Kyle Wade, who call the tool 'the video editing version of

Cursor .'

Mosaic | Agentic AI Video Editing Platform
https://mosaic.so/

Adish Jain on Reinventing Video Editing with Mosaic's AI Agents | Interview | AIPressRoom
https://aipressroom.com/adish-jain-mosaic-interview/

The idea for Mosaic began when Jayne and Wade were trying to create a YouTube video for themselves. They were trying to count the number of Cybertrucks driving around town, but they found it extremely time-consuming and stressful to manually edit specific scenes from the vast amount of footage they had to film.

Existing commonly used editing software such as Adobe Premiere Pro and DaVinci Resolve have complex functions, and even simple operations require searching, resulting in a high learning curve. Mosaic was developed to eliminate the complexity and learning curve of video editing, allowing creators to focus on storytelling.

Mosaic's greatest feature is that, unlike traditional editing software, it uses a 'canvas' where functional blocks (nodes) are connected to build an editing flow. By placing tiles with functions such as 'cut,' 'subtitle generation,' and 'B-roll insertion' on this canvas and connecting them with lines, an AI agent can automatically perform a series of tasks.



Mosaic's AI agent also has multimodal capabilities, meaning that the AI can understand both the visual and auditory information in a video, just like a human. For example, if you instruct the AI agent to 'highlight scenes with dogs in them,' the AI will recognize the dog in the video and automatically apply appropriate effects and subtitles. Unlike conventional AI tools that simply cut out silent sections, Mosaic's AI agent is capable of advanced editing that understands the content and context of the video.



Mosaic is designed for beginners and professionals alike. It integrates the latest generative AI models, including Google Veo 3, Kling 1.6, and Gemini 2.5 Pro, to deliver high-quality video and audio synthesis. Furthermore, the edited data can be exported in XML format to DaVinci Resolve, Adobe Premiere Pro, and Final Cut Pro, allowing for seamless integration with existing software. After creating the overall edit, Mosaic allows for fine-tuning in your preferred software.

It also has a function to simultaneously generate different video patterns from the same material and perform A/B testing, allowing businesses and marketers to efficiently find the most effective video content.



Jain said that in the future, he aims to create an AI agent with a creative sense called 'ARGO (A Really Good Opinionated agent).' This will not just follow instructions, but will also make suggestions and decisions on its own to create a 'good video,' acting as a kind of partner that seamlessly supports the entire process from filming to distribution.

At the time of writing, Mosaic is available in beta , but to fully utilize its AI capabilities, you need to sign up for a subscription plan starting from $25 (about 2,800 yen) per month or $500 (about 75,000 yen) per year.

Related Posts:

in AI,   Video,   Software,   Web Application, Posted by log1i_yk