Video Editor

Generate, review, and manage AI-generated videos for your shots using the Video Editor.

Overview

The Video Editor is a dedicated page for generating and managing video content for individual shots. Using AI video models, you can transform your keyframes into dynamic video sequences, compare multiple video variations, and select the best one for your production.

Key Concepts

• Shot Video: An AI-generated video based on various inputs (keyframes, prompts, reference images, or existing videos)
• Generation Mode: The method used to create the video — one of eight modes selected from the Mode dropdown (see the Video Generation Modes section below)
• First Frame / Keyframe: The starting image for first-frame video generation
• Video Prompt: A text description guiding the AI to create the desired motion and action
• Video Model: The AI model used to generate the video (filtered based on selected mode)
• Selection Status: Mark each video as Selected, Discarded, or leave it Pending to designate the official shot video

When to Use the Video Editor

The Video Editor is where you:

• Generate shot videos: Create AI-powered videos using multiple generation modes
• Transform existing videos: Edit, upscale, or restyle videos with Edit Video and Upscale Video modes
• Compare variations: Generate multiple versions with different prompts, models, or modes
• Review and select: Watch videos and select the best one for production
• Extract frames: Capture specific frames from videos to use as keyframes for other shots
• Download videos: Export videos for external use or archival

---

Accessing the Video Editor

From the Video Manager

1. Open your project and go to the Videos section — from the Project Sections list on the Project Hub, or from the section-name dropdown in the breadcrumb at the top of any project page. This opens the Video Manager.
2. In the Video Manager, shots are grouped by scene as cards. Each card shows a video preview or a Never Generated badge, plus a Jump to Shot number field if you want to skip ahead.
3. On the shot you want to work on, click the Edit video in Keyframe icon to open the Video Editor for that shot.

Editing Shot Details Instead

Each Video Manager card also has an Edit Shot button (pencil icon). Instead of opening the Video Editor, it opens the Edit Shot drawer, where you can adjust the shot's metadata: assignments, shot direction (size, angle, movement, subject count), subjects, transition, director's notes, scene property overrides (weather, atmosphere, lighting notes, location, art medium), the first frame (with a button into the Keyframe Editor), and dialog lines.

Navigation

• Use the breadcrumb section menu to return to the Video Manager or move to any other part of your project.
• The header displays your project name and type.

---

Understanding the Interface

The Video Editor has three main areas:

Left Sidebar

The sidebar contains controls for generating and managing videos:

Main Video Player

The central area displays the video player:

• Video Display: Full-size video playback area
• Player Controls: Play, pause, timeline scrubbing, volume, playback speed, fullscreen
• Empty State: Displays guidance message when no video is selected

Bottom Thumbnail Strip

The thumbnail strip displays all generated videos for this shot:

• Each thumbnail shows the video's prompt, the model id used (for example, fal-ai/veo3.1/image-to-video), and the duration (for example, "Duration: 5.0s")
• Filter tabs at the top of the strip — Selected, Pending, and Discarded — each with a count, let you narrow the strip to videos in that state
• Hover a thumbnail for its per-video actions: Mark as selected, Mark as discarded, and Media info
• Click a thumbnail to load that video in the player
• Scroll horizontally to access more videos
• The video loaded in the player is visually highlighted

---

Video Generation Modes

The Video Editor supports eight generation modes, each suited for a different creative workflow. Select the mode from the Mode dropdown in the sidebar.

Mode Overview

Mode Selection

When you select a mode:

1. The sidebar updates to show the inputs relevant to that mode
2. The Video Model dropdown filters to only the models that support that mode
3. Your prompt text is preserved when switching modes
4. Mode-specific inputs (keyframe selectors, source-video pickers, and so on) appear or hide as appropriate

---

First Frame Only Mode (I2V)

The default and most common mode. Generate a video from a single keyframe image.

When to Use First Frame Only

• You have a selected keyframe and want to animate it
• Standard shot video generation
• Most straightforward workflow

How to Use First Frame Only

1. Select First Frame Only from the Mode dropdown
2. Verify or select the First Frame (your shot's keyframe)
   • Click "Select First Frame" to choose a different keyframe
   • Use the "Selected Only" toggle to filter by selection status
3. Enter a Video Prompt describing the desired motion
4. Select a Video Model from the dropdown (displays First Frame Only-compatible models)
5. Click Generate

---

Text to Video Mode (T2V)

Generate a video purely from a text prompt without any reference images.

When to Use Text to Video

• Abstract concepts that don't have visual references
• Establishing shots or background scenes
• When no keyframe exists yet
• Exploring creative concepts before creating keyframes

How to Use Text to Video

1. Select Text to Video from the Mode dropdown
2. The first frame selector is hidden (not needed)
3. Enter a detailed Video Prompt describing the entire scene and motion
4. Select a Video Model from the dropdown (displays T2V-compatible models)
5. Click Generate

Tip: T2V prompts should be more descriptive since there's no visual reference. Include details about the scene, characters, lighting, and motion.

---

First-Last Frame Mode (KF2V)

Generate a video that smoothly transitions between a first frame and last frame, creating interpolated motion.

When to Use First-Last Frame

• Cinematic transitions between two compositions
• Controlled motion from point A to point B
• When you know exactly how a shot should start and end
• Camera movements or character repositioning

How to Use First-Last Frame

1. Select First-Last Frame from the Mode dropdown
2. Select the First Frame:
   • Click "Select First Frame" or use "Use Keyframe" to load the shot's selected keyframe
   • Toggle "Selected Only" to filter available images
3. Select the Last Frame:
   • Click "Select Last Frame" to choose the ending image
   • This can be from any shot's keyframes or uploaded images
4. Enter a Video Prompt describing the transition motion
5. Select a Video Model from the dropdown (displays KF2V-compatible models)
6. Click Generate

Note: Both first frame and last frame are required. The AI interpolates motion between them based on your prompt.

---

Edit Video Mode (V2V)

Transform an existing video with style transfer, enhancement, or re-interpretation. Optionally sync to dialog audio.

When to Use Edit Video

• Style transfer (change the visual style of existing video)
• Re-interpret a video with different characteristics
• Sync video to dialog audio for lip-sync or timing
• Fix issues in an existing video while preserving motion

How to Use Edit Video

1. Select Edit Video from the Mode dropdown
2. Select the video you want to transform in the thumbnail strip; it becomes the Source Video shown in the sidebar
   • If no video is selected, the sidebar prompts you to pick one from the thumbnail strip
3. (Optional) Turn on Sync with Dialog if dialog audio exists for this shot
   • When enabled, the resulting video duration matches the shot's dialog length
4. Enter a Video Prompt describing the transformation
5. Select a Video Model from the dropdown (displays Edit Video-compatible models)
6. Click Generate

Sync with Dialog

When dialog audio is available and sync is enabled:

• The generated video length matches the dialog duration
• Useful for ensuring lip-sync or scene timing
• The source video is stretched/compressed to match

---

Upscale Video Mode (UPSCALE)

Enhance video resolution and quality without changing content. Unlike other modes, Upscale Video does not require a prompt.

When to Use Upscale Video

• Increase video resolution for final delivery
• Enhance quality of a video you're happy with
• Prepare videos for high-resolution output
• Post-production quality enhancement

How to Use Upscale Video

1. Select Upscale Video from the Mode dropdown
2. Select the video you want to upscale in the thumbnail strip; it becomes the source shown in the sidebar
   • If no videos exist, you'll see a message to generate one first
3. Select a Video Model from the dropdown (displays Upscale Video-compatible models)
4. Configure Upscale Settings (if available for the selected model):
   • Scale factor (e.g., 2x, 4x)
   • Target resolution
   • Target FPS
5. Click Generate

Note: No prompt is required for Upscale Video mode. The AI enhances the video using algorithmic upscaling.

---

References to Video Mode (R2V)

Generate a video using reference images from your project's assets (cast members, locations, props) for character and scene consistency.

When to Use References to Video

• Maintain character consistency across shots
• Use specific asset appearances in video generation
• Complex scenes requiring multiple visual references
• When you need the AI to understand specific character designs

How to Use References to Video

1. Select References to Video from the Mode dropdown
2. Click to open the Reference Image Selector:
   • Browse images from Cast, Locations, Props, and other assets
   • Toggle "Selected Only" to filter by selection status
   • Select 1 to N reference images (typically up to 4)
3. Enter a Video Prompt describing the scene and motion
4. Select a Video Model from the dropdown (displays References to Video-compatible models)
5. Click Generate

Reference Image Tips

• Select character reference images for consistent appearance
• Include location references for background consistency
• More references can help the AI understand your scene better
• At least one reference image is required

---

Motion Control Mode

Drive the motion of a subject using an existing video as the motion source. The AI applies the movement from a "driving" video to a subject you supply.

When to Use Motion Control

• Transfer the motion of one video onto a specific subject
• Reuse a movement you like with a different character or look
• Direct precise, repeatable motion from an existing clip

How to Use Motion Control

1. Select Motion Control from the Mode dropdown
2. Select the video that supplies the motion in the thumbnail strip; it becomes the Driving Video shown in the sidebar
   • If no video is selected, the sidebar prompts you to pick one from the thumbnail strip to use as the motion source
3. Choose a Subject Keyframe — the image whose subject will follow the driving video's motion
4. (Optional) Enter a Video Prompt to guide the result
5. Select a Video Model from the dropdown (displays Motion Control-compatible models)
6. Click Generate

---

Multi-Shot Video Mode

Generate a run of consecutive shots as a single continuous video, with a prompt for each shot in the range.

When to Use Multi-Shot Video

• Produce a continuous sequence across several shots in one pass
• Keep motion and style consistent from one shot to the next

How to Use Multi-Shot Video

1. Select Multi-Shot Video from the Mode dropdown
2. Click Pick ending shot to define the range of shots to include; the panel then shows the shot range and shot count (for example, "Shots 3–6 (4 shots)"). Use Change ending shot to adjust the range.
3. Under Per-shot prompts, enter a prompt for each shot in the range.
4. If the selected model supports per-shot durations, set a duration for each shot; the panel shows the total.
5. Select a Video Model from the dropdown (displays Multi-Shot Video-compatible models)
6. Click Generate

---

Media Selection Filtering

When selecting keyframes, videos, or reference images, the Video Editor provides selection filtering.

Selected Only Toggle

All media selection dialogs include a "Selected Only" toggle:

• Enabled (default): Displays only selected media for production-quality inputs
• Disabled: Displays all media including unselected items

Why Use Selection Filtering

• Ensures only reviewed content is used for generation
• Maintains production quality standards
• Helps focus on finalized assets
• Follows the same pattern as the Keyframe Editor

When No Selected Media Exists

If "Selected Only" is enabled but no selected media exists:

• A message displays: "No selected [media type] available"
• You can toggle the filter to see unselected options
• Consider selecting assets before video generation

---

Video Player Controls

When a video is loaded in the player, you have full playback control:

The player fills as much of the available space as possible for optimal viewing.

---

Managing Videos

Loading a Video

Click any thumbnail in the strip to:

• Load that video in the main player
• Populate the prompt field with that video's prompt

The video loaded in the player is visually highlighted.

Selection Workflow

Each video has a selection state that tracks which version is the official shot video:

Setting a video's state:

1. Hover the video's thumbnail in the strip
2. Click Mark as selected to make it the official video, or Mark as discarded to set it aside
3. Use the Media info action to inspect a video's details
4. The change saves immediately, and the strip's Selected / Pending / Discarded filter tabs update their counts

One Selected Video Per Shot:

Only one video per shot can be selected at a time. When you select a video:

• Any previously selected video is automatically reset to pending
• This happens atomically — no confirmation required
• The filter-tab counts update immediately

Extracting Frames

The Extract Frame feature lets you save a specific video frame as a keyframe for any shot:

1. Play the video and pause at the desired frame
2. Click Extract Frame in the sidebar
3. A dialog opens displaying the current frame
4. Select which shot should receive this frame as its keyframe
5. Confirm the extraction

This is useful for:

• Creating keyframes for shots that follow the current one
• Capturing the perfect moment from a video
• Building visual continuity across scenes
• Providing the "last frame" for First-Last Frame mode on the next shot

Downloading Videos

To download a video for external use:

1. Load the video in the player
2. Click Download in the sidebar
3. The video file downloads to your browser's download location

Use this to:

• Archive videos locally
• Share with stakeholders outside the platform
• Use in external editing applications

Uploading a Video

Use Upload Video in the sidebar to bring an existing video file into the shot instead of generating one. Uploaded videos appear in the thumbnail strip alongside generated ones and can be selected or discarded the same way.

---

Approved Shots Are Locked

Once a shot is Approved, its generation controls are locked to protect the approved result. On an approved shot, the Mode dropdown, prompt, model selector, Generate button, New Video, and Upload Video are all disabled.

---

Multiple Video Variations

The Video Editor supports generating multiple video variations per shot, allowing you to experiment and compare before committing to a final choice.

Why Generate Multiple Variations?

• Different prompts: Try various descriptions to get the perfect action
• Different models: Compare output from different AI models
• Different modes: Use First Frame Only, Text to Video, or References to Video to explore approaches
• Iterative refinement: Build on what works, discard what doesn't
• Stakeholder options: Present multiple versions for review

Browsing Variations

Use the thumbnail strip to:

• Scroll through all generated videos
• Quickly compare thumbnails
• Click to preview any version
• See selection status at a glance

Comparing Videos

To compare videos effectively:

1. Click a thumbnail to load it in the player
2. Watch the video, noting what works and what doesn't
3. Click another thumbnail to compare
4. Repeat until you find the best version
5. Select your final choice

---

Video Models

The Video Editor offers various AI video models, each with different characteristics. Models are filtered based on your selected generation mode.

Model Filtering by Mode

When you select a generation mode, only the models that support that mode appear. Each of the eight modes — First Frame Only, First-Last Frame, References to Video, Text to Video, Edit Video, Motion Control, Multi-Shot Video, and Upscale Video — shows its own compatible set of models.

Selecting a Model

1. First select your generation mode
2. Click the Video Model dropdown in the sidebar
3. Browse available models (filtered for compatibility)
4. Note the pricing (displayed per second)
5. Select your preferred model

Model Considerations

When choosing a model, consider:

• Mode compatibility: Only models supporting your selected mode are displayed
• Cost: Different models have different per-second rates
• Duration: Some models produce longer videos
• Style: Models may have different visual characteristics
• Speed: Generation time varies by model
• Quality: Higher-cost models often produce better results

Model Categories

Common model types include:

---

Assignment and Workflow

Assignee

The assignee section displays who is responsible for this shot's video:

• View the current assignee
• Change assignment (if you have permission)
• Unassigned shots display "Unassigned"

Status

Track workflow progress with status:

---

Tips and Best Practices

Mode Selection Tips

• Start with First Frame Only if you have a selected keyframe
• Use Text to Video for abstract or establishing shots without visual reference
• Use First-Last Frame when you know exactly how the shot should start and end
• Use Edit Video to restyle or sync existing videos
• Use Motion Control to drive a subject's movement from an existing clip
• Use Multi-Shot Video to generate a run of consecutive shots as one continuous video
• Use Upscale Video as a final polish step
• Use References to Video for character consistency across scenes

Prompt Writing Tips

• Be specific about motion: Describe exactly what should move and how
• Include timing: Mention speed (slowly, quickly) when relevant
• Describe camera: Specify camera movement if desired (pan, zoom, static)
• Reference the scene: Mention key elements from the shot
• Keep it concise: Clear, focused prompts often work better than lengthy ones
• More detail for T2V: Without visual reference, prompts need more scene description

Workflow Tips

• Start with the keyframe: Ensure your first frame is exactly what you want
• Generate multiple versions: Try different prompts, models, and modes
• Review systematically: Watch each video fully before deciding
• Select early: Set selection status so downstream processes know which video to use
• Use Extract Frame: Capture great moments for use in other shots
• Upscale last: Enhance quality after selecting your preferred video

Performance Tips

• Be patient: Video generation can take time
• Check costs: Monitor credit usage, especially with expensive models
• Use faster models for drafts: Save premium models for final versions
• Keep the strip tidy: Mark unwanted takes as Discarded and use the Discarded filter tab to keep them out of your way

---

Common Questions

Why is the Generate button disabled?

Requirements vary by mode. If the shot is Approved, the Generate button and other controls are also disabled — see the Approved Shots Are Locked section below.

Why don't I see certain models?

Models are filtered by generation mode. If you don't see a model:

• Check that you've selected the correct mode
• That model may not support your selected mode
• Try a different mode to see different models

Why can't I use Edit Video or Upscale Video mode?

Edit Video and Upscale Video work on an existing video:

• Generate a video using another mode first (First Frame Only, Text to Video, First-Last Frame, or References to Video)
• Then use Edit Video or Upscale Video to transform or enhance it

Why does video generation take so long?

Video generation is computationally intensive. Time depends on:

• The selected model
• Video duration
• Generation mode complexity
• Server load

Some models offer "fast" or "turbo" variants for quicker results.

Can I edit a generated video?

Generated videos cannot be edited within the Video Editor. To make changes:

• Generate a new video with a modified prompt
• Use Edit Video mode to transform the video
• Assemble and trim shots in the project's Timeline, then use Timeline Export to finish in external software

What happens if video generation fails?

If generation fails:

• An error message appears in the player area
• The loading skeleton is removed
• You can retry by clicking Generate again
• Credits are not charged for failed generations

How do credits work in the Video Editor?

Can I switch modes and keep my prompt?

Yes! Your prompt text is preserved when switching between modes. This makes it easy to try the same prompt with different generation approaches.

What's the difference between First Frame Only and References to Video?

Can I have multiple selected videos?

No, each shot can have only one selected video. Selecting a new video automatically resets the previous one to pending status.

---

Next Steps

• Timeline Export - Export your assembled timeline to an editing application
• Keyframe Editor - Create and refine keyframes for video generation
• Dialog Editor - Add voice audio to your shots
• Subscriptions & Credits - Understand credit usage for video generation