New AI Features in DaVinci Resolve
Version note: This guide describes the AI workflow available in Resolve Studio 20.2. Features can vary between the free and Studio editions and may change with point releases, so check Help > About DaVinci Resolve and the official release notes before relying on a feature in a client project.
Artificial intelligence in DaVinci Resolve is most useful when it removes repetitive work without taking control away from the editor. The latest Resolve workflow brings AI-assisted tools into editing, tracking, audio cleanup, color, and delivery.
1. IntelliTrack and AI-powered tracking
IntelliTrack helps follow people and objects through a shot so you can attach effects, masks, titles, or color adjustments more quickly. It is especially useful for screen replacements, subject callouts, selective grading, and motion graphics.
Start with a clean frame, check the track across the whole shot, and correct problem areas manually. AI tracking is a time-saver, not a reason to skip quality control.
2. Text-based editing
Text-based editing can turn a transcript into an editing surface. Search for a phrase, select the part you need, and build a rough cut from spoken content before refining the picture in the timeline.
This is valuable for interviews, podcasts, tutorials, and social clips. Always review the source audio because automatic transcripts can misunderstand names, technical terms, accents, and overlapping speakers.
3. Voice Isolation and dialogue cleanup
Voice Isolation uses AI-assisted processing to reduce unwanted background noise and make dialogue easier to understand. It can help with room tone, traffic, fans, and other distractions that are difficult to remove by hand.
Use the lowest strength that gives you a clean result. Heavy processing can create metallic artifacts, so compare the treated voice with the original before exporting.
4. Smart Reframe for different platforms
Smart Reframe helps keep a subject inside the frame when converting a project for vertical, square, or other social formats. It can speed up the first pass when one source edit needs several aspect ratios.
After reframing, review every cut and title. A human check is still important when the subject changes position, multiple people share the frame, or the shot uses intentional negative space.
5. Magic Mask and object selection
Magic Mask can help isolate people, objects, or regions for targeted color and effects work. AI-assisted selections are useful for a fast starting point, especially in complex shots where a basic shape would take too long.
Refine edges, inspect motion blur, and check hair or transparent objects at full resolution. Small corrections make the difference between a convincing composite and an obvious cutout.
6. Scene detection and faster organization
Scene analysis tools can help locate cuts and organize long recordings into workable sections. This gives editors a quicker overview of interviews, presentations, and archive footage before the detailed edit begins.
Use the result as an assistant-generated starting point. Confirm every detected cut before deleting or rearranging source material.
7. AI tools for color and finishing
Resolve's color workflow includes AI-assisted tools that can help identify faces, track subjects, match shots, and speed up repetitive corrections. These features are most effective when the editor establishes a consistent creative direction first.
AI can help with the technical pass, but it does not know the intended mood, story, or client reference. Keep creative decisions in your hands.
8. AI audio organization
AI-assisted audio tools can help editors find dialogue, separate useful material from noise, and prepare a cleaner starting point for the Fairlight page. Searchable transcripts and speaker information are particularly helpful when a project contains several interviews or long recordings.
Use this organization pass before making detailed audio decisions. Listen to important lines in context, preserve room tone where it supports continuity, and avoid treating every background sound as a problem. A natural recording usually feels better than an aggressively processed one.
9. Object removal and cleanup
AI-assisted object removal can help clean up a distracting mark, unwanted sign, or temporary item in the background. It works best when the object is reasonably separated from the subject and the surrounding texture gives Resolve enough information to rebuild the area.
Short shots with simple camera movement are a good place to begin. Review the result frame by frame around movement, reflections, shadows, and overlapping objects. For a client delivery, keep the original clip and the cleaned version available so you can quickly compare the change.
10. AI tools for social content
Editors creating short-form content can combine transcription, smart reframing, subject tracking, and voice cleanup into one repeatable process. Start by finding the strongest spoken moments, build a tight vertical or square rough cut, then refine captions, pacing, graphics, and music manually.
Do not let automatic framing decide every composition. A speaker may need space for a headline, a product may need room for a call to action, and a reaction shot may work better when the face is not centered. AI can create the first version; editorial intent should create the final version.
What is actually new for editors?
The important change is not one magic button. It is the way several assistant tools now cover the repetitive stages of post-production. An editor can move from transcript to rough cut, from rough cut to tracked graphics, and from noisy dialogue to a usable mix with fewer manual steps between each decision.
That speed is most valuable when it gives you more time for judgment. You can compare more takes, test a stronger opening, clean up a difficult interview, or prepare multiple deliverables without cutting corners. The goal is not to make every project look automatic. The goal is to make the invisible work faster so the visible work gets more attention.
Free versus Studio features
Not every AI feature is available in every edition of DaVinci Resolve. Some tools require Resolve Studio, while others may have different limits or behavior depending on the installed release. Before planning a workflow around a feature, open the tool in your own installation and confirm that it is available.
For teams, record the Resolve version and edition in the project notes. This simple detail prevents a common handoff problem: a project opens correctly on one machine but loses an AI-assisted effect or adjustment on another. Consistent version information is part of a professional workflow.
A practical AI workflow
For a new interview project, begin with transcription or scene organization. Mark the best answers and build a rough timeline. Use voice cleanup only after choosing the takes, then apply tracking, Magic Mask, or reframing where the story needs visual support. Finish with a human review of dialogue, color, titles, and exports.
Save milestones such as 01_transcript_selects, 02_ai_cleanup, and 03_final_review. Clear versions make experimentation safer and help another editor understand what an AI tool changed. They also make it easier to return to the original creative decision when an automatic result is not right.
How to use AI without losing control
- Duplicate the timeline before applying an AI-assisted operation.
- Test the feature on a short representative section first.
- Inspect edges, dialogue artifacts, tracking errors, and reframing choices.
- Save a version with clear notes so the change can be reversed.
- Confirm the feature is available in your Resolve edition and version.
Final thoughts
The best new AI features in DaVinci Resolve are practical assistants: they help editors search, track, isolate, clean, and reframe faster. The strongest results come from combining that speed with human judgment, careful review, and a reliable versioning workflow.