Runway

Runway is a creative platform offering multiple AI video and image generation tools, including Gen-1, Gen-2, and Gen-3 models. It's popular for video-to-video, image-to-video, and text-to-video workflows.

Platform Overview

• Provider: Runway Research Inc.
• Specialty: Video generation and editing
• Access: Free tier with watermark, paid plans for commercial use
• Supported Output Types: Video, Image

Lazer Detection

URL Patterns

The extension detects Runway on:

• https://runwayml.com
• https://app.runwayml.com

What Is Auto-Detected

When you generate with Runway, Lazer extracts:

• Prompt - From prompt textarea or generation card
• Model - Gen-1, Gen-2, Gen-3, or specific tool name
• Duration - From duration selector (video)
• Aspect Ratio - From ratio buttons
• Resolution - From quality settings
• Output URL - Video/image src from results
• Thumbnail - Video poster or image preview

Detection Quality

• Prompt extraction: High - Dedicated prompt fields
• Settings extraction: Medium - Some settings hidden in advanced options
• Output extraction: High - Reliable media URL detection

Capture Workflow

Text-to-Video Workflow (Gen-3)

1. Open Runway Gen-3 tool
2. Enter your prompt
3. Set duration and aspect ratio
4. Click "Generate"
5. Wait for generation (can take 30-60 seconds)
6. Video preview appears
7. Open Lazer extension
8. Detection banner shows "Runway"
9. Click "Auto Fill"
10. Review fields (especially model and duration)
11. Click "Save Capture"

Image-to-Video Workflow (Gen-2)

1. Open Gen-2 tool
2. Upload reference image
3. Enter motion prompt
4. Set duration
5. Generate
6. Wait for result
7. Capture as above

Multiple Generations

For iterative work:

1. Generate first variation
2. Capture it
3. Adjust prompt or settings
4. Generate again
5. Click "Refresh Detection"
6. Capture new variation

Capture Tips

Best Practices

• Wait for full quality - Initial preview may be compressed
• Note generation mode - Gen-1, Gen-2, or Gen-3 have different capabilities
• Check watermark - Free tier adds visible watermark
• Include motion keywords - For image-to-video, motion description is critical

Watermark on Free Tier

Runway free tier includes:

• Visible "Runway" watermark on videos
• Watermark appears in corner or center
• Removed with paid subscription
• Important for rights tracking

Common Issues

Prompt not detected in multi-step tools:

• Some Runway tools have multi-step workflows
• Prompt may be on a previous screen
• Manually copy prompt before advancing

Duration not captured:

• Duration selector may be hidden after generation
• Manually note video length and add to metadata

Model version unclear:

• Runway has Gen-1, Gen-2, Gen-3, plus specialized tools
• Detector attempts to identify from UI
• Manually verify model name if needed

Rights and Licensing

Free vs Paid Plans

Runway licensing varies by plan:

Free Plan

• Limited credits per month
• Visible watermark on all outputs
• Non-commercial use recommended
• Outputs subject to Runway terms

Creator/Pro Plans

• No watermark
• Full commercial license
• Higher resolution and quality
• You own your outputs

Commercial Use

Per Runway's terms:

• Free tier: Use at your own risk, watermark present
• Paid plans: Full commercial rights, no attribution required
• Training data: Cannot use outputs to train competing models

Lazer Rights State

The extension sets rightsState to:

• RESTRICTED - If watermark detected (inferred from free plan)
• LICENSED - If authenticated with paid plan (inferred)
• UNKNOWN - Default if plan unclear

Manually set rights state in web app based on your Runway subscription.

Known Limitations

Multi-Step Workflows

Some Runway tools have multi-step UIs:

• Prompt entered on step 1
• Generation happens on step 2
• Detection works best on results page

Async Generation

Runway generations can take time:

• Progress bar shows generation status
• Don't capture until "Complete" status
• May need to refresh page to see final result

Tool Variety

Runway has many specialized tools:

• Gen-3 Alpha (text-to-video)
• Gen-2 (image-to-video, video-to-video)
• Inpainting, Expand Image, etc.

Each tool may have slightly different UI. Detector uses common patterns.

Metadata Fields

Runway-specific metadata:

{
  "modelName": "Gen-3 Alpha",
  "duration": 10,
  "aspectRatio": "16:9",
  "resolution": "1080p",
  "watermarkPresent": false,
  "platform": "runway"
}

Example Capture

A typical Runway capture in Lazer:

• Platform: runway
• Asset Type: VIDEO
• Prompt: "A drone shot rising above a foggy forest at sunrise"
• Model: Gen-3 Alpha
• Output URL: https://storage.runwayml.com/video_xyz789.mp4
• Thumbnail: https://storage.runwayml.com/thumb_xyz789.jpg
• Metadata: {"modelName": "Gen-3 Alpha", "duration": 10, "aspectRatio": "16:9", "watermarkPresent": false}
• Status: GENERATED
• Rights State: LICENSED

Troubleshooting

Detection Not Working

• Verify you're on a results/generation page
• Wait for generation to fully complete
• Click "Refresh Detection"
• Grant host permissions for runwayml.com

Output URL Expired

• Runway uses temporary storage URLs
• Extension syncs immediately to avoid expiry
• Verify successful sync in Queue mode

Watermark Status Unclear

• Free tier always has watermark
• Paid plans remove watermark
• Visually inspect video to verify
• Set rights state to RESTRICTED if watermark present

Model Name Generic

• Detector captures model from UI badges
• Some tools show generic "Runway" instead of Gen-1/2/3
• Manually note specific model in tags or notes field

Multi-Step Prompt Lost

• Copy prompt text before advancing steps
• Paste into capture form manually if detection fails

Related Resources

• Runway Homepage
• Runway Pricing
• Gen-3 Alpha Announcement
• Runway Terms of Service

Next Steps

• Compare with Sora and Veo
• Understanding rights and watermarks
• Capture workflow guide