Capture Mode

Capture Mode is the primary interface for saving generated assets from AI platforms to your Lazer scenes. It combines auto-detection, ruthless defaults, and one-click saving to minimize friction in high-volume capture workflows.

Overview

When you generate content on an AI platform, Capture Mode:

1. Auto-detects the platform, prompt, settings, and outputs
2. Pre-fills form fields with detected data
3. Applies context defaults (last project, scene, asset type)
4. Enables one-click save to queue
5. Syncs to your Lazer web app in the background

Interface Components

Context Selectors

At the top of Capture Mode, you'll find three dropdown selectors:

• Project - Choose which film project this asset belongs to
• Scene - Choose which scene this asset belongs to
• Shot - (Optional) Choose which specific shot this asset belongs to. Defaults to "(Scene-level)" if not specified.

These selectors remember your last selection and apply it as the default for subsequent captures.

When a shot is selected, the capture is associated with that specific shot. When "(Scene-level)" is selected, the asset is attached to the scene without a specific shot.

Selecting a scene automatically loads that scene's existing assets into History mode for reference.

Detection Banner

The detection banner displays real-time platform and output detection:

• Platform Indicator - Shows detected platform name and logo
• Detection Dot - Green (detected), red (not detected)
• Asset Type - Shows detected output type (image, video, audio)
• Auto Fill Button - Populates form with detected data
• Refresh Button - Re-runs detection on current page

Platform Detection States

• Green dot + platform name - Successfully detected, ready to capture
• Red dot + "No platform detected" - Not on a supported platform or detection failed
• Yellow dot + "Detecting..." - Detection in progress

Capture Form

The form is organized into essential and advanced fields:

Essential Fields (Always Visible)

Platform

• Auto-filled from detection
• Dropdown to manually override if needed
• Synced from your Lazer platform list

Prompt

• The text prompt used to generate this asset
• Auto-filled from page content
• Actions:
  • Apply - Inject prompt back into platform page
  • Refine - AI-powered prompt enhancement (requires AI config)
  • Clear - Reset field

Output URL

• Direct link to the generated media file
• Auto-detected from video/image elements
• Used by Lazer to fetch and store the asset

Type

• Asset type: Image, Video, Audio, Music, Voice, Narration, Storyboard, Other
• Defaults to detected type
• Controls how asset is displayed in web app

Status

• Asset status: Draft, Generated, Selected
• Controls review workflow in web app
• Defaults to "Draft"

Advanced Fields (Expandable)

Click "More fields" to reveal:

Title

• Optional descriptive name for this variant
• Useful for tracking iterations ("take-2", "director-notes")

Negative Prompt

• Things to avoid in generation
• Auto-filled if platform supports negative prompts

Model

• AI model name/version (e.g., "Veo-3", "Sora-turbo")
• Auto-filled from platform UI when visible

External ID

• Platform's job ID or generation ID
• Useful for tracing back to platform

Source URL

• The platform page URL where this was generated
• Auto-filled with current tab URL

Thumbnail URL

• Preview image URL (for video assets)
• Auto-detected from video poster attribute

Tags

• Comma-separated tags (e.g., "hero, approved, pass-2")
• Useful for filtering and search

Metadata JSON

• Additional structured data in JSON format
• Example: {"seed": 42, "steps": 50}
• Auto-populated with detected settings

Notes

• Free-form text notes about this version

Actions

Clear All

• Resets the entire form to defaults
• Preserves project and scene selection

Save Capture

• Adds item to sync queue
• Optimistically marks as "queued"
• Background worker syncs to Lazer API
• Shows success or failure in Queue mode

Auto-Fill Workflow

The recommended workflow leverages auto-detection:

1. Navigate to an AI platform and generate content
2. Wait for generation to complete
3. Detection banner turns green with platform name
4. Click "Auto Fill" in the detection banner
5. Review auto-populated fields
6. Adjust Type or Status if needed
7. Click "Save Capture"

Auto-fill extracts prompt, settings, output URLs, and metadata from the page DOM.

Ruthless Defaults

To enable one-click capture, the extension remembers and applies:

Last Active Context

• Project - Last selected project
• Scene - Last selected scene
• Asset Type - Last used type
• Status - Last used status

These defaults persist across browser sessions.

Auto-Detected Values

• Platform - Detected from current URL
• Prompt - Extracted from prompt input or generation card
• Output URL - Extracted from video/image elements
• Model Name - Extracted from UI badges
• Aspect Ratio - Extracted from settings
• Duration - Extracted from video controls
• Source URL - Current tab URL

Fallback Behavior

If detection fails for a field, the extension:

• Leaves the field empty
• Does not block capture (only Project, Scene, Platform, and Prompt are required)
• Allows manual entry

Batch Capture

For platforms that generate multiple variations at once:

1. Click "Auto Fill" to detect the first output
2. Review and adjust fields
3. Click "Save Capture"
4. Click "Refresh Detection" to detect the next output
5. Click "Auto Fill" again
6. Repeat until all variants are captured

Future versions may support true batch capture with a "Save All" button.

Manual Capture

If auto-detection fails, you can manually capture:

1. Copy the prompt text from the platform
2. Paste into the Prompt field
3. Right-click the generated output and copy image/video address
4. Paste into Output URL field
5. Manually select Platform from dropdown
6. Fill any other desired fields
7. Click "Save Capture"

Field Validation

The extension validates required fields before queueing:

• Project - Required
• Scene - Required
• Platform - Required
• Prompt - Required (minimum 1 character)
• Asset Type - Required

If validation fails, the extension highlights the missing field and shows an error message.

Note: Output URL is optional. Captures can be saved without a media URL to record prompts and metadata for later generation.

Capture Tips

For Video Assets

• Ensure the video is fully loaded before capturing
• Check that thumbnailUrl is detected (used for grid previews)
• Capture duration and resolution metadata when available

For Image Assets

• Wait for high-resolution image to load (not just thumbnail)
• Verify the URL points to the full-size image, not a preview
• Some platforms serve images via CDN with expiring URLs

For Audio Assets

• Capture waveform or spectrogram thumbnail if available
• Include duration in metadata
• Note licensing (especially for music platforms)

For Prompt Reuse

• Keep prompts clean and reusable (avoid one-off references)
• Use prompt refinement to standardize formatting
• Tag prompts that work well for later reuse

Integration with Other Modes

History Mode

After capturing an asset:

• It appears in History mode for the scene
• You can click it to load its data back into the form
• Useful for iterating on a winning prompt

Queue Mode

After clicking "Save Capture":

• Item is added to the sync queue
• A prompt package is automatically created alongside the asset version
• Background worker attempts upload
• Check Queue mode to verify success
• Failed items remain in queue for retry (up to 10 attempts for server errors; client errors like 401, 400, 404 fail immediately)

Troubleshooting

Platform Not Detected

• Verify you're on a supported platform
• Check that you're on a generation or results page
• Click the refresh icon to retry detection
• Grant host permissions if prompted

Prompt Not Extracted

• Some platforms hide prompts after generation
• Manually copy/paste the prompt before generating
• Check "More fields" for negative prompt

Output URL Not Found

• Ensure content is fully loaded
• Look for download buttons or links
• Right-click the media and copy address manually
• Some platforms require user interaction to reveal URLs

Save Fails Immediately

• Check that all required fields are filled
• Verify project and scene are selected
• Ensure Output URL is a valid HTTP/HTTPS URL
• Check browser console for validation errors

Item Stuck in Queue

• Switch to Queue mode to check sync status
• Click "Sync Now" to retry
• Clear the item if it fails after 10 retries
• Verify your API token is valid and not expired
• Check for client errors (401, 400, 404) which fail immediately and require manual intervention

Next Steps

• Reuse captured prompts
• Manage the sync queue
• Platform detection reference