# AI video editing prompts: 45 for Premiere Pro, DaVinci Resolve and CapCut

45 prompts and step-by-step workflows for cutting, captions and cleanup in Premiere Pro, DaVinci Resolve and CapCut, including Voyager-assisted handoffs.

Author: Ravi Parikh

Canonical: https://voyager.so/blog/ai-video-editing-prompts

Published: 2026-11-17T11:09:00-08:00
Updated: 2026-11-17T11:09:00-08:00

AI video editing prompts can help turn raw footage into a structured cut, accurate captions and a cleaner final export. For 2026, the practical challenge is not just generating instructions but getting those instructions into editable production workflows. This guide covers 45 prompts and step-by-step workflows for [Premiere Pro](https://voyager.so/resources/apps/premiere-pro), [DaVinci Resolve](https://voyager.so/resources/apps/davinci-resolve), [CapCut](https://voyager.so/resources/apps/capcut) and Voyager's agent-assisted workflows. It separates instructions you can give an agent from controls you must use inside an editor, with a focus on handoffs that remain useful after the first AI-generated cut.

[Voyager](https://voyager.so/) is a desktop AI-agent harness that connects supported creative applications, editing skills, models and revisable files. It can coordinate a workflow around Premiere or Resolve, while CapCut tasks usually require prepared media and manual finishing. The distinction between rendered videos and [editable video projects](/blog/what-editable-means-for-ai-video) matters when revisions are inevitable. For a broader map of editing approaches, see [AI video editor workflows](/blog/best-ai-video-editors).

**The short version:** Start with the outcome and the editable handoff. Voyager can use agent instructions and its creative skills to prepare footage, captions, audio and timelines for supported editing environments. Premiere workflows can use importable XML timelines and supported control routes; DaVinci Resolve Studio 21.1+ has a native MCP route; CapCut generally requires prepared clips or SRT files followed by manual editing. None of these routes removes the need to review timing, captions, effects or exports.

**Key takeaways**

* **Voyager connects editing stages:** Its agent and creative skills can coordinate generation, rough cuts, captions and [editable project handoffs](/blog/what-editable-means-for-ai-video). The exact result depends on the chosen editor and integration route.
* **Premiere workflows need a supported handoff:** Voyager can prepare FCP7 XML timelines; edits to existing sequences have additional control requirements. Review [Premiere MCP setup](/blog/premiere-pro-mcp) before assuming live project control.
* **Resolve MCP depends on the edition:** DaVinci Resolve Studio 21.1+ offers a native MCP server, while free-edition workflows have different limits. Check [Resolve MCP workflows](/blog/davinci-resolve-mcp) before running agent actions.

The 45 entries below include 15 natural-language prompts and 30 interface workflows. A prompt is a brief for an agent or a compatible tool, not a guarantee that every menu action is remotely controllable. Required software licenses, supported operating systems, integration setup and generation charges still apply.

Sources: [Adobe Premiere Help](https://helpx.adobe.com/premiere/desktop/whats-new/whats-new.html), [Blackmagic Design’s Resolve manual](https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf) and [CapCut Help](https://www.capcut.com/help/how-to-recognise-subtitles), read 17 November 2026. Voyager, the desktop app our team builds, supports the editable handoffs discussed here.

## How AI video editing prompts work

AI video editing prompts describe the editorial result you want, while the connected agent or editor determines which actions it can actually perform. In Voyager, you can describe the task and ask for a timeline, subtitle file, organized assets or changes in a supported desktop application. Transcription, analysis and editing still depend on the available tools, permissions and footage. An [agent workflow comparison](/blog/best-ai-agent-platforms-for-video-and-audio) helps distinguish creative orchestration from editor-specific features.

### Two approaches to AI editing

* **Prompt-based editing** uses written instructions to guide AI decisions. You describe what you want (remove filler words, preserve pauses before key points, match a specific pace), and a capable agent translates that intent into supported actions or an editable handoff. Voyager can orchestrate this route using its creative tools and supported editor integrations. A prompt alone does not grant an agent access to an application's timeline.
* **Tool-based editing** uses an editor's built-in features and controls. You select a tool such as Magic Mask or Voice Isolation, adjust the available parameters and inspect the result. An agent may be able to invoke selected controls through MCP, scripting or computer control, but other steps must still be done manually. The instructions below label these workflows separately from copy-paste prompts.

### The prompt formula

A useful editing prompt describes the desired outcome, relevant footage, pacing, constraints, and elements that should remain unchanged. These details give the editing agent clearer instructions while leaving room to handle variations in the footage.

A basic prompt structure includes:

* **Content type:** YouTube tutorial, podcast, vlog, interview
* **Primary action:** Remove filler words, cut silences, generate captions
* **Parameters:** Pause threshold (2 seconds), tone (professional, casual), exceptions
* **Keep list:** What should remain unchanged

## Cutting prompts by platform

Cutting workflows vary depending on the editor, available AI features, and integration method. The following prompts and instructions cover rough cuts, silence removal, transcript editing, and footage selection across Premiere Pro, DaVinci Resolve, and CapCut, including workflows that Voyager can support through editable handoffs and compatible agent connections.

### Premiere Pro cutting prompts

For Voyager-assisted Premiere work, ask the agent for a structured rough cut and an importable FCP7 XML timeline. Existing-project edits depend on the supported control route and app version. [Premiere Pro 27.0](https://helpx.adobe.com/premiere/desktop/whats-new/whats-new.html) also introduced a native AI Assistant, so the prompts below describe editorial intent rather than one third-party plugin's guaranteed commands.

#### Prompt 1: Basic talking-head cleanup

```text
Edit this talking-head YouTube tutorial into a clear, professional rough cut. Remove unnecessary filler words such as "um" and "uh," repeated sentences, false starts, and pauses longer than 2 seconds.

Preserve intentional pauses before important explanations, natural breathing room, complete sentences, and transitions between topics. Avoid cutting the beginning or end of meaningful words.

Maintain a conversational but focused pace throughout the video. Keep the original topic order and do not remove technical explanations or examples that help viewers understand the subject.

Prepare the edited timeline for review, and flag any uncertain cuts or transcript sections that may need manual attention.
```

#### Prompt 2: Podcast episode with natural flow

```text
Edit this podcast episode to remove unnecessary verbal stumbles while preserving the conversational tone. Remove "um," "uh," repeated words, and false starts that interrupt the flow.

Keep pauses before key points, natural conversation rhythm between speakers, thoughtful moments that indicate reflection, and laughter or authentic reactions. Maintain the casual, friendly tone throughout.

Do not remove pauses shorter than 2 seconds unless they occur mid-sentence. Preserve complete questions and answers, topic transitions, and speaker overlaps that feel natural.

Flag any sections where the conversation is difficult to follow or where speaker identification is unclear. Prepare the timeline for a review pass focusing on pacing and speaker balance.
```

#### Prompt 3: Interview with multiple speakers

```text
Edit this interview to maintain professional pacing while preserving thoughtful exchanges. Remove long silences between speaker turns (longer than 3 seconds), repeated questions or answers, and obvious false starts.

Keep reaction pauses where one speaker is processing what another said, brief silences that add weight to important statements, natural overlaps that show engagement, and transitions between topics. Maintain a documentary-style professional tone.

Identify each speaker clearly throughout the timeline. Do not remove pauses that indicate the interviewee is considering their response. Preserve complete thoughts even if they take multiple sentences to express.

Flag any sections where audio quality differs significantly, where speaker attribution is ambiguous, or where a cut might alter the meaning of a statement. Prepare the timeline for content review and speaker balance check.
```

#### Prompt 4: Fast-paced social content

```text
Edit this short-form video for maximum energy and retention. Remove all pauses longer than 0.5 seconds, repeated phrases, any verbal stumbles, and slow sections that do not advance the content.

Keep punch lines and their setup timing, emphasis moments where repetition is intentional, transitions between visual demonstrations, and the original content order. Maintain high energy pacing suitable for TikTok or Reels.

Do not remove pauses that are part of a comedic beat or emphasis pattern. Preserve clear audio on every word. Cut aggressively to maintain momentum without making the edit feel jarring.

Flag any sections where the edit feels too abrupt or where audio cuts mid-word. Verify that the final timeline maintains continuity and that no important content is lost. Target a runtime under 60 seconds unless the content requires more.
```

#### Prompt 5: Educational course content

```text
Edit this online course lecture for student comprehension while maintaining instructional clarity. Remove verbal stumbles, false starts, long pauses (over 4 seconds), and repeated explanations unless the repetition serves a teaching purpose.

Keep deliberate pauses after key concepts for student note-taking (2-3 seconds), complete explanations even if they take time, transitions between topics with clear signposting, and any on-screen demonstrations or examples being discussed.

Maintain instructional pacing that allows learners to follow along. Do not rush through complex topics. Preserve the order of concepts as presented. Keep any verbal cues that reference course materials or assignments.

Flag any sections where an explanation is incomplete, where timing between verbal and visual elements is misaligned, or where the edit might confuse the learning sequence. Prepare the timeline for educational review focusing on clarity and completeness.
```

After reviewing the selected takes, preserve a copy of the source project and import the agent's compatible timeline into Premiere. Verify media relinking and cuts before applying further effects. Live adjustments to an existing sequence require an integration route that supports them.

### DaVinci Resolve workflow instructions

DaVinci Resolve has built-in editing controls and, in Studio 21.1 or later, a native MCP server for compatible agents. Voyager can use edition-aware Resolve routes; the free edition has narrower app-control support, including platform restrictions. The following entries are manual interface workflows, not commands guaranteed to execute through MCP.

#### Instruction 6: Auto-split long recordings

```text
DaVinci Resolve Studio > Edit page > Select the clip containing existing scene changes > Timeline > Detect Scene Cuts > Review the automatically detected edit points > Correct any inaccurate cuts manually
```

Scene detection identifies existing visual cuts in footage, rather than automatically segmenting a continuous recording by topic or speech. This is useful for screen recordings, exported projects, or multicamera footage.

#### Instruction 7: Extract highlights from webinar

```text
Use Scene Cut Detection to segment footage > Mark in/out on best segments > Create subclips > Arrange on new timeline
```

#### Instruction 8: Remove dead AIr in podcast

```text
DaVinci Resolve Studio > Fairlight page > Select the relevant audio range > Right-click > Remove Silence > Adjust threshold, minimum silence duration, and fade settings > Review detected sections > Apply
```

[Resolve Studio's AI Remove Silence tool](https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf) identifies silent sections for removal. Unlike an audio gate, which attenuates quiet sound without shortening the timeline, this workflow is designed to remove silent sections from the audio edit.

#### Instruction 9: Batch process B-roll selection

```text
Media Pool > Organize clips using metadata, keywords, and Smart Bins > In DaVinci Resolve Studio 21, use AI IntelliSearch to find footage containing specific people, objects, or dialogue > Review the matching clips > Add suitable B-roll selections to the timeline
```

[Smart Bins](https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf) organize clips based on metadata rules, while AI IntelliSearch provides content-based searching.

![Blackmagic Design’s AI IntelliSearch section showing content search in DaVinci Resolve’s Media Pool](https://cdn.moda.app/voyager/blog/ai-video-editing-prompts/resolve-intellisearch-02bdbd5792c16238.webp)

IntelliSearch finds faces, objects and dialogue; Smart Bins instead use metadata rules. Documentation checked 17 November 2026.

### CapCut transcript-based cutting

CapCut offers [transcript-based editing](https://www.capcut.com/resource/edit-video-with-text) for removing unwanted dialogue and pauses. Voyager can prepare clips, audio and subtitle files for CapCut. There is no native CapCut MCP route, although permission-based computer control can operate its desktop interface on supported systems. For background on unsupported or unofficial automation routes, see [CapCut MCP limits](/blog/capcut-mcp). Feature availability depends on platform, app version, region and plan.

#### Workflow 10: Filler word removal

```text
Import footage into CapCut Desktop > Open Menu > Layout > Transcript-based editing > Generate the transcript > Locate filler words such as "um" and "uh" > Review each occurrence > Delete unnecessary words from the transcript and linked timeline
```

#### Workflow 11: Pause-based cutting

```text
Open Transcript-based editing > Review detected speech gaps and pauses > Identify pauses longer than approximately 1.5 seconds > Remove unnecessary sections > Preserve intentional pauses and transitions > Preview the edited timeline
```

#### Workflow 12: Repeated take removal

```text
Open Transcript-based editing > Search for repeated phrases or restarts > Compare the available takes manually > Keep the clearest complete take > Remove unwanted repetitions > Check transitions and audio continuity
```

#### Prompt 13: Voyager asset handoff for CapCut

```text
Using the source footage, identify the strongest takes for a YouTube rough cut. Remove unnecessary filler words, repeated statements and pauses longer than 2 seconds only where a supported tool can make reliable cuts. Do not change the meaning of the speaker's statements.

Prepare an ordered set of trimmed clips, a cut list with source timecodes, an SRT caption file and separate audio assets where available. Preserve the original footage and identify any missing or uncertain segments.

Package these assets for manual import and finishing in CapCut. Do not claim that you edited a CapCut timeline unless the project was actually opened and the result checked. Flag any work that remains manual.
```

Note: This is a Voyager preparation prompt, not a direct CapCut MCP command. A CapCut project should be reviewed inside the app after import; optional computer control requires explicit permission, and compatibility with third-party project formats should not be assumed.

## Caption prompts by platform

AI-assisted captioning can help generate subtitles, improve readability, and prepare content for different platforms. These prompts cover caption styling, speaker identification, subtitle exports, and accessibility, with practical applications in [AI feature tutorials](/blog/feature-tutorial-video) and Voyager-supported editing workflows.

### Premiere Pro caption prompts

These examples are caption styling briefs for a Voyager-assisted workflow or a compatible editing tool. Voyager can help prepare speech text and subtitle files, but specific fonts, animations and layouts need to be implemented and verified in the destination editor. For a completed captioned-video example, see this [feature tutorial workflow](/blog/feature-tutorial-video).

#### Prompt 14: Brand-colored captions

```text
Generate captions for this video with the following specifications:

Font: Montserrat Bold
Primary text color: #FF5722 (brand orange)
Highlight color: #FFFFFF (white)
Position: Center-bottom, safe for mobile viewing
Animation: Word-by-word reveal

Apply consistent styling throughout the video. Ensure caption timing is synchronized with speech. Use the highlight color for emphasis on key terms and action words. Keep captions legible at mobile viewing sizes.

Review generated captions for proper nouns, technical terms, and brand names that may need manual correction. Verify that caption position does not obscure important visual elements.
```

#### Prompt 15: YouTube tutorial captions

```text
Generate captions for this multi-speaker tutorial video with clear speaker identification:

Include speaker labels for each person
Use contrasting colors for different speakers (assign automatically or use: Speaker 1 blue #2196F3, Speaker 2 green #4CAF50)
Maximum 2 lines per caption block
Minimum 32pt font size for mobile viewing
Position: Lower third, avoiding channel logo and subscribe button area

Include full punctuation for readability. Time captions to match speaker changes. Do not overlap captions when speakers talk simultaneously; prioritize the primary speaker.

Review captions for speaker attribution accuracy. Verify that color coding is consistent throughout. Check that technical terminology is spelled correctly and that acronyms are properly formatted.
```

#### Prompt 16: Social media captions

```text
Generate captions optimized for social media short-form content:

Large centered text (48-60pt minimum)
Pop-in animation on keywords and important phrases
Bold, high-contrast font suitable for small screens
Auto-highlight: action words, numbers, and emotional peaks

Use a single caption style throughout for consistency. Time animations to emphasize key moments without distracting from the content. Ensure captions are readable in both light and dark viewing environments.

Keep caption blocks short (5-8 words maximum). Review for proper emphasis placement. Verify that animated highlights do not obscure the speaker's face or important visual elements.
```

#### Prompt 17: Accessibility-focused captions

```text
Generate captions following accessibility best practices:

Include full punctuation and proper grammar
Add speaker labels for all speakers
Describe relevant sound effects in brackets [door closes], [phone rings]
High contrast: white text on black background with semi-transparent backing
Minimum 28pt font size
Position: Lower third, avoiding overlap with other UI elements

Time captions to allow sufficient reading time (minimum 2 seconds per caption block, adjusted for word count). Break longer sentences across multiple caption blocks for readability.

Review captions for completeness, including all spoken dialogue and relevant audio information. Verify that sound effect descriptions are accurate and helpful. Check that caption timing allows comfortable reading without rushing.
```

### DaVinci Resolve caption workflows

DaVinci Resolve 21 introduced expanded subtitle and captioning features. These workflows use the Captions panel in the Inspector and Fusion effects.

#### Instruction 18: Auto-generate subtitles

```text
DaVinci Resolve Studio > Edit page > Open the target timeline > Timeline > AI Tools > Create Subtitles from Audio > Select language and caption settings > Create > Review subtitle text and timing in the Inspector
```

#### Instruction 19: Styled lower thirds

```text
Fusion page > Text+ node > Connect to MediaOut > Add Follower modifier for word-by-word reveal > Apply to timeline as Fusion Composition
```

#### Instruction 20: Multi-language subtitle export

```text
DaVinci Resolve Studio > Generate subtitles in the original spoken language > Review and export the subtitle track as .srt > Translate the .srt externally using an appropriate translation tool > Review the translated text and timing > Import each translated .srt into Resolve on a separate subtitle track > Export the required language versions
```

[Resolve 21's documented subtitle features](https://documents.blackmagicdesign.com/UserManuals/DaVinci_Resolve_21_Reference_Manual.pdf) do not establish a native automatic translation command. Translation requires a separate workflow, and generated subtitles should be reviewed before delivery.

#### Instruction 21: Broadcast-safe captions

```text
Deliver page > Render Settings > Subtitle Settings > Enable Export Subtitle > Select As embedded captions where supported > Choose a compatible output format and caption standard > Verify placement and inspect the exported file
```

### CapCut caption generation

CapCut offers automatic captions and template-based styling. Caption availability and usage limits vary by [platform](https://www.capcut.com/help/how-to-recognise-subtitles), region and [plan](https://www.capcut.com/help/capcut-standard-vs-pro).

![CapCut Help’s desktop auto-caption instructions with the Captions and Auto Captions controls](https://cdn.moda.app/voyager/blog/ai-video-editing-prompts/capcut-captions-127e2db0d9d3fe7b.webp)

The desktop instructions identify the controls behind these caption workflows; availability still depends on region, version and plan. Documentation checked 17 November 2026.

#### Workflow 22: Auto captions with templates

```text
Select clip in timeline > Captions > Auto captions > Choose language > Generate > Apply an available caption template > Customize colors and fonts as needed > Preview and adjust timing
```

#### Workflow 23: Word-level highlighting

```text
Generate captions as above > Edit caption style panel > Choose an available word-by-word highlight style or animation > Set highlight color and animation style > Adjust timing offset if captions are slightly out of sync > Preview across the full timeline
```

#### Workflow 24: Multi-speaker identification

```text
Generate captions > Enable speaker detection (if available in your version) > Assign colors per speaker for visual distinction > Position speaker name labels > Review speaker attribution accuracy > Correct any misattributed dialogue manually
```

#### Workflow 25: Vertical video captions

```text
Generate captions for 9:16 aspect ratio content > Center position for visibility > Increase font size to 48pt minimum for mobile readability > Add text shadow or background for legibility > Check safe area to avoid overlap with profile icons or UI elements > Preview on a phone-sized viewport
```

#### Workflow 26: Karaoke-style lyrics

```text
CapCut mobile app > Import a video containing clear vocals > Open Text > Select Auto Lyrics where available > Generate synchronized lyrics > Review recognition accuracy and timing > Customize available text styles and animations > Preview and export
```

CapCut documents [Auto Lyrics as a mobile feature](https://www.capcut.com/help/entrance-to-auto-lyrics). It is not supported on the desktop or web versions in its platform-specific help documentation.

## Cleanup prompts by platform

Video cleanup involves improving audio clarity, correcting colors, stabilizing footage, and refining visual elements before export. The following workflows explain how to apply these adjustments in creative applications while preserving [editable video projects](/blog/what-editable-means-for-ai-video) for further revisions.

### Premiere Pro cleanup workflows

Premiere Pro provides built-in cleanup tools. Voyager can coordinate compatible assets and timeline handoffs, while audio effects and finishing controls must be checked in the installed version.

#### Workflow 27: Audio enhancement pass

```text
Select the dialogue clip in the timeline > Properties panel > Audio > Enhance Audio > Choose the available Enhance Speech model > Adjust dialogue enhancement, volume, and reverb reduction > Preview the result > Check final loudness separately before export
```

[Adobe's Enhance Speech V1](https://helpx.adobe.com/premiere/desktop/add-audio-effects/adjust-volume-and-levels/enhance-speech.html) runs locally, while V2 uses cloud processing and generative credits.

#### Workflow 28: Color match across clips

```text
Lumetri Color > Color Wheels & Match > Enable Comparison View > Select a reference frame and the current clip > Apply Match > Review and fine-tune the color adjustments > Repeat for additional clips
```

#### Workflow 29: Auto-reframe for social

```text
Select the sequence in the Project panel > Sequence > Auto Reframe Sequence > Choose the target aspect ratio or custom dimensions > Select an appropriate motion-tracking preset > Enable clip nesting if needed > Create the reframed sequence > Review framing and adjust keyframes manually
```

#### Workflow 30: Background noise reduction

```text
Select the audio clip > Apply the DeNoise audio effect > Open Effect Controls > Adjust the Amount slider gradually > Preview the dialogue > Reduce the effect strength if speech sounds unnatural
```

### DaVinci Resolve AI cleanup tools

DaVinci Resolve offers several [cleanup controls](https://documents.blackmagicdesign.com/SupportNotes/DaVinci_Resolve_Studio_21_Features.pdf), with advanced AI features depending on the Studio edition. Voyager can coordinate supported Resolve operations, but edition-specific effects and manual review still apply; see the [Resolve app reference](https://voyager.so/resources/apps/davinci-resolve). Do not assume that an MCP connection unlocks Studio-only features.

#### Instruction 31: Voice Isolation

```text
Fairlight page > Select dialogue track > Mixer Track FX (or Inspector) > Voice Isolation > Enable > Adjust isolation strength (start at 50%, increase as needed)
```

This can reduce background noise during dialogue, but strong settings may affect voice quality. Compare the processed result with the original recording.

#### Instruction 32: Auto Color correction

```text
Color page > Select clip > Color menu > Auto Color (or the A button in Primaries) > Review > Fine-tune with primaries if needed
```

#### Instruction 33: Shot Match across footage

```text
Color page > Select target clips > Right-click the reference clip > Shot Match to This Clip > Verify skin tones and highlight consistency
```

#### Instruction 34: Magic Mask subject isolation

```text
Color page > Magic Mask panel > Click the subject > Track forward/backward > Invert for background adjustments > Apply secondary correction
```

#### Instruction 35: Dialogue Leveler for consistent audio

```text
Fairlight page > Select the dialogue track > Enable Dialogue Leveler in the available audio processing controls > Adjust and preview dialogue consistency > Apply loudness normalization separately for the required delivery target > Configure music ducking independently using the appropriate track automation or sidechain controls
```

Dialogue Leveler, loudness normalization, and music ducking are separate processing tasks.

### CapCut cleanup features

CapCut includes manual cleanup features used in short-form finishing. These controls are listed for editors who import Voyager-prepared media and finish the cut in CapCut.

#### Workflow 36: One-click quality enhancement

```text
Select clip > Basic panel > Enhance Quality (upscales resolution) > Adjust panel > Auto Adjust (balances color, brightness, contrast)
```

#### Workflow 37: Background removal

```text
Select clip > Video > Remove BG > Auto removal > Review edges > Use Customized removal brushes for hair/complex edges > Add replacement background
```

#### Workflow 38: Noise reduction for mobile footage

```text
Select clip > Video > Basic > Enable Reduce image noise > Adjust the strength > Preview at full size > Balance noise reduction against detail preservation
```

#### Workflow 39: Stabilization for handheld shots

```text
Select clip > Basic > Stabilize > Choose mode (Recommended, Minimum cut, Most stable) > Preview > Crop compensation if needed
```

#### Workflow 40: Auto-sync to music beats

```text
Add music track > Select music > Auto-Mark Beats > Choose Beats 1 (strong only) or Beats 2 (all beats) > Snap clip edges to markers
```

## Refining prompts for better results

Clear prompts reduce ambiguity, especially when an agent must hand work between assets, timeline formats and desktop applications. For mixed-source footage, [joining AI clips](/blog/ffmpeg-join-ai-video-clips) also requires checking format and synchronization. Detail is useful only when the available integration can perform the requested step.

### Add exceptions to preserve intent

Instead of "remove all pauses," specify thresholds and exceptions:

```text
Remove pauses longer than 2 seconds, but keep pauses before key points
```

```text
Cut filler words except when used as transitions between topics
```

```text
Remove repeated phrases, but keep intentional emphasis repetition
```

### Specify content type and tone

Different content types have different pacing expectations:

* Podcasts benefit from conversational pauses that tutorials do not need
* Social content typically requires faster pacing than educational material
* Interviews should preserve thoughtful pauses that indicate reflection

### Include technical parameters

When possible, include specific values:

* Pause threshold in seconds
* Target loudness in LUFS
* Font size in points
* Color values in hex codes

### Build a prompt library

Create saved prompts for recurring content formats, including the source files, destination editor, output format, limits and review checklist. In Voyager, reusable skills can preserve workflow instructions across projects; the [video agent brief](/blog/how-to-brief-an-ai-video-agent) is a useful starting structure. Record any manual correction needed so future prompts describe actual constraints rather than assuming a successful automated edit.

## Voyager agent prompts for cross-app editing

The next five prompts focus on the part that individual editor commands cannot cover alone: planning, preparing and revising work across tools. In Voyager, use the connected agent, supported editing skills, available MCP servers and editable file formats; confirm the route before requesting a change to an existing project. These prompts extend the cutting, caption and cleanup examples into practical production handoffs.

### Cross-app agent prompts

#### Prompt 41: Build a Premiere-compatible rough cut

```text
Inspect the supplied footage, transcript and edit brief. Select the strongest complete takes and prepare a concise rough cut with pauses preserved before key explanations. Keep the source order unless the brief explicitly requires a new structure. Create a Premiere-compatible FCP7 XML timeline with linked source-media references where supported, plus a cut list with timecodes. Do not overwrite the existing Premiere project. Flag unsupported effects, missing files and any edit that needs manual review after import.
```

#### Prompt 42: Prepare a Resolve timeline through a supported route

```text
Check which DaVinci Resolve edition, version, operating system and integration route are available before making changes. On a supported Studio 21.1+ MCP connection, inspect the project and duplicate the target timeline before making supported changes to clip order and markers. For trims, rebuild a new timeline from source ranges if the integration cannot cut existing clips. If direct editing is unavailable, prepare an importable timeline or documented manual steps instead. Do not attempt Studio-only operations on the free edition. Summarize the changes and list the shots that still need review.
```

#### Prompt 43: Prepare captions and audio for handoff

```text
Transcribe the approved dialogue and create an SRT file with accurate timecodes, punctuation and speaker labels where reliable. Keep caption blocks readable for a 9:16 mobile version, but do not assume a particular editor's animation preset exists. Prepare a separate music and dialogue mix plan, including a target loudness only if the delivery specification supplies one. Preserve the editable caption file, flag unclear words, and explain which timing or styling tasks remain manual.
```

#### Prompt 44: Coordinate motion graphics and video

```text
Using the provided brand guide and approved copy, plan three editable lower-thirds and one closing title card for the video. If an After Effects connection with the required scripting or computer-control permissions is available, create them as revisable text and shape layers. Otherwise prepare a layer and keyframe specification for manual construction. Keep typography, safe areas and timing consistent with the cut. Provide an inspection checklist and do not describe an unrendered preview as final output.
```

#### Prompt 45: Create revisable social cut variants

```text
From the approved master timeline, prepare 16:9, 9:16 and 1:1 edit variants for a product demo. Reframe each shot for the primary subject, update caption placement for the platform safe area and preserve all essential on-screen information. Use Voyager Video or a supported editor handoff when available, and leave the timelines editable. Identify any shots that cannot be reframed without losing meaning, and request approval before generating replacement material or exporting the final versions.
```

## Using Voyager with Premiere Pro, DaVinci Resolve and CapCut

Voyager works as an agent harness around the production rather than as a substitute for every control inside every editor. It can plan a cut, generate or prepare assets, and preserve revisable project handoffs; the exact editing permissions depend on the app. The built-in Voyager Video, Design and DAW engines also provide a route when an external desktop editor is unnecessary.

### Voyager Video, Design and DAW

* Voyager Video can create editable timelines and export footage or supported interchange formats for further editing.
* Voyager Design can prepare editable title and motion assets; more complex After Effects work depends on the app and its control route. For authentic screens, see the [real UI workflow](/blog/real-ui-product-video); for integration requirements, see [After Effects MCP](/blog/after-effects-mcp).
* Voyager DAW can help arrange narration, music and sound cues, with the final mix checked against the destination's delivery requirements.

### Premiere Pro: editable timeline handoffs

* Prepare a rough cut and hand it off as an importable FCP7 XML timeline, checking media paths and unsupported effects after import.
* Use a supported bridge or computer-control route for operations on an existing sequence; a prompt by itself does not open or edit a Premiere project.
* Keep the native project and source media intact, and review the imported result before export.

### DaVinci Resolve: MCP and edition limits

* DaVinci Resolve Studio 21.1+ includes a native MCP route for supported operations. Confirm the connection and duplicate the timeline before edits.
* Free-edition commands have narrower support, including macOS requirements for in-app import and editing; Windows may require a prepared timeline and manual import.
* Check advanced effects, subtitle exports and render settings in Resolve instead of assuming agent access bypasses edition restrictions.

### CapCut: prepared assets, manual finishing

* Use Voyager to organize footage, music, voiceovers, cut lists and SRT captions for import. A [character launch example](/blog/talking-character-launch-video) demonstrates timing captions and sound against short scenes.
* Finish transcript cuts, caption styles and effects in CapCut itself, manually or through permission-based computer control; this is not a native CapCut MCP integration.
* Reopen and preview the final timeline, especially after using version-dependent auto-caption or stabilization tools. For changes to generated shots, the [character replacement workflow](/blog/ai-video-character-replacement) shows what to compare.

## Final verdict: where Voyager fits

Voyager brings AI video editing prompts into a connected creative production workflow by combining AI agents, supported desktop applications, generative models, and [editable project files](/blog/what-editable-means-for-ai-video). Its [Premiere Pro workflows](/blog/premiere-pro-mcp) and [DaVinci Resolve MCP integration](/blog/davinci-resolve-mcp) help creators prepare structured rough cuts, coordinate editing tasks, and maintain revisable timelines across production stages.

Voyager's built-in Video, Design, and DAW editors further support video editing, motion graphics, audio production, and creative revisions within an agent-assisted environment. For creators, video editors, and marketing teams looking to turn natural-language instructions into practical production workflows, Voyager offers a flexible way to connect AI-assisted creation with professional editing and reusable project assets.

## Frequently asked questions

### What is the difference between prompt-based and tool-based AI editing?

Prompt-based editing uses natural-language instructions to describe a task, while tool-based editing uses preset features with adjustable settings. Voyager can coordinate supported agent actions and editable handoffs; Resolve's Magic Mask and Voice Isolation are examples of editor controls. Both approaches always require checking what the connected application actually supports.

### Do AI editing prompts work on footage without speech?

Many cutting prompts use transcripts and work best with dialogue, but agents can organize B-roll, titles and music-led edits. For nonverbal footage, use scene detection, beat markers or manual editorial cut points. Voyager can prepare structured assets for a [feature tutorial example](/blog/feature-tutorial-video), while color correction and stabilization still require visual checks.

### How much does AI video editing with Voyager cost?

Voyager is free to download, including its Video, Design and DAW editors. Its free plan provides limited managed usage, while [paid subscriptions](https://voyager.so/pricing) starting at US$19/month (checked 17 November 2026) add broader generation access; your own agent or provider keys may incur external charges. Creative applications such as Premiere Pro or Resolve Studio require separate licenses. Editing time, generation credits, model selection and output requirements affect the full cost. Check current plan terms before committing to a production budget.

### Can AI-generated captions handle technical terminology and proper nouns?

AI captions can mishandle names and technical terms. Review generated captions before publishing, especially when a product demo contains interface labels or names of people and places. Voyager can prepare [editable subtitle files](/blog/what-editable-means-for-ai-video), but captions still need checks for spelling, synchronization and readability.

### Should I run AI cleanup tools before or after cutting?

Start with a duplicate of the source project and remove unwanted takes or filler words before generating final captions. Then review timing, adjust color and audio, and export from the destination editor. If Voyager prepares an [editable timeline](/blog/what-editable-means-for-ai-video), verify its media references and the changes after importing it. Applying cleanup earlier may be useful for noisy source audio, but avoid processing footage that will be discarded. Keep the original files and an editable project so you can revise uncertain cuts without degrading the original source media.


## Voyager is a desktop AI agent for creative work

Describe what you want to make; Voyager plans it, uses your apps and tools, and keeps everything editable.

[Sign up](/sign-up)