What to know

  • Project Frame Forward edits entire videos using changes made in a single frame.
  • The AI automatically applies edits like object removal or addition across all frames.
  • The tool enables context-aware modifications, such as adding objects that interact naturally with scene elements.
  • Adobe introduced this at its 2025 Max conference as part of experimental AI video editing features.

Adobe introduced Project Frame Forward at their 2025 Max conference as an innovative AI-powered system designed to revolutionize video editing by making it more like photo editing. Instead of editing every frame individually, users modify only the first frame, and the AI applies that change to the whole video clip. For example, a person removed from the initial frame will automatically be removed from subsequent frames with a natural-looking background fill.

How Project Frame Forward works

The AI takes two inputs - the original video and the edited first frame. It analyzes the video to understand the scene and then intelligently propagates the initial frame's modifications throughout the video, maintaining consistency and realism.

This system is capable of removing unwanted elements or adding new ones based on sketches and AI prompts. Added objects are context-sensitive, meaning they will interact with or reflect the movement of existing elements in the scene.

Alongside Project Frame Forward, Adobe also showcased:

  • Project Light Touch: Uses generative AI to manipulate light sources in photos and videos, allowing users to change lighting direction, simulate lamps turning on, or create dynamic lighting effects.
  • Project Clean Take: Focuses on audio editing, enabling changes to speech pronunciation and delivery using AI prompts.

Adobe's Project Frame Forward promises to improve workflows for creators by automating traditionally tedious processes.