Adobe’s annual Max conference is underway, and Premiere is getting a lengthy list of updates spanning color grading, audio cleanup and generative AI. Adobe is also expanding its AI assistants and previewing connections that let third-party AI agents work directly in Premiere projects. As with the company’s photography announcements, some of these tools have already appeared in beta, though there are more updates for the video side of things.

Color Mode leaves beta

Adobe

Color Mode is becoming generally available after Adobe introduced it in public beta in April. It gives Premiere a dedicated color-grading environment, with a simplified user interface separate from the typical timeline view. It’s designed to make color grading more approachable for video editors, without requiring the specialized expertise of a professional colorist.

As with the initial version, it features “Style Presets,” which Adobe says are more flexible than applying a LUT. They are editable, stackable collections of adjustments that let you adjust different aspects of the look, rather than forcing you to accept a fully-baked transformation

With the presets, editors can see how a look affects the sequence at a glance, making it easy to see if everything matches. If there are individual clips or groups of clips that need additional adjustments, the editor can select those to fine-tune the look. It should be a useful tool when most of an edit looks consistent, but a handful of shots need extra attention.

Generative Remove tackles unwanted objects in video

Premiere’s Generative Remove is entering beta, giving editors a way to remove unwanted objects from footage and fill the resulting space with generated detail.

In Adobe’s briefing, the tool used automated masks to identify the object before removing it nondestructively. The idea is straightforward enough: get rid of the distraction without rebuilding the surrounding scene manually or using other programs. Of course, how convincingly it handles complicated movement and backgrounds will be something to assess in testing.

More generation happens inside the timeline

Adobe is also expanding the Generative Media Tool, which lets editors create assets directly in Premiere rather than generate them elsewhere and import the results.

The tool already supported video and sound-effect generation. During the briefing, Adobe demonstrated image generation as another option, alongside music and soundscapes. The generated assets become part of the edit, keeping more of the workflow inside the timeline.

Music and soundscape generation are not entirely new announcements: Adobe introduced both in beta in September. Generate Soundscape analyzes footage to create accompanying audio, while Generate Music offers controls for shaping a track to fit the project. During the briefing, Soundscape produced four tracks, giving the editor separate elements to work with rather than a single combined recording.

Audio cleanup gets more control

Adobe

Premiere’s audio updates include improved Enhance Audio, Separate Crosstalk and Separate Sound Source, which Adobe says help clean up recordings and separate their components.

In the keynote, Adobe demonstrated independently adjusting or muting elements within a single recording. For example, it showed a clip of someone speaking on a boat with loud engine noise and music playing. It used the Enhance Audio tool to quickly bring down all of the background noise, so that you could hear the speaker much more clearly. It could be particularly useful for field recordings, where dialogue, background noise and other sounds often arrive together whether you want them or not.

Additionally, the Separate Crosstalk tool helps when audio clips have multiple speakers talking over each other. The example provided in the keynote involved an interview in which the subject started talking before the interviewer finished the question. With Separate Crosstalk, it automatically put each speaker on their own track, so they could mute the interviewer and keep the subject’s audio clear.

Some of this has a history, too. Adobe announced expanded audio-separation capabilities in September. And at last year’s Max Sneaks presentation, its experimental Project Clean Take demonstrated separating a recording into voice, background sounds and music.

Premiere’s AI Assistant takes on more work

Adobe is expanding Premiere’s AI Assistant, a conversational tool powered by its Creative Agent. Rather than just explaining how to do something, it can carry out editing tasks in response to natural-language requests.

The assistant first entered public beta in June. It remains in beta, with Adobe highlighting expanded capabilities such as applying effects across every clip in a sequence, tracking masked subjects and adjusting color and exposure across an entire sequence.

During the briefing, Adobe also showed workflows such as synchronizing cameras and audio for multicamera clips and flagging shots that may have been missed during color grading.

Editors can save repeatable instructions as custom “skills,” then call them up with a slash command instead of explaining the same workflow each time. Adobe frames it as writing down how you like a task handled so you don’t have to start from scratch on the next project.

Adobe’s Creative Agent is also coming to Premiere mobile in private beta. It will let users turn footage into editable first cuts, request changes using natural language and continue refining the results themselves.

Third-party agents get a route into Premiere

Adobe is separately previewing local Model Context Protocol, or MCP, integrations that let third-party agents work directly in native Premiere, Photoshop, Illustrator and After Effects projects.

For Premiere users, the important part is that bins and sequences remain editable. An agent can make changes inside the project, and the editor can step in to inspect or revise them rather than receive only a finished export. The integrations are a preview, not a generally available feature at this point.

That is separate from Premiere’s built-in AI Assistant. One gives users an assistant inside Adobe’s app; the other opens a route for their preferred third-party agent to work with the project.

Follow DPReview to ensure our independent journalism reaches your feeds.