version: "1.0.1" name: higgsfield-video-editing description: Edit and enhance videos using Higgsfield.ai editing tools. ALWAYS enable unlimited mode + visible browser with captcha alerts. BATCH PROCESSING - edit multiple videos in sequence. Use when user wants to (1) extend video duration, (2) add effects, (3) loop videos, (4) enhance quality, (5) add subtitles. Uses agent-browser CLI.

Higgsfield Video Editing

⚠️ ALWAYS CLOSE BROWSER WHEN DONE

# ✅ ALWAYS close browser when finished
agent-browser close

IMPORTANT: Always run agent-browser close when done to properly release the headed mode session.

⚠️ CRITICAL: BROWSER MUST BE VISIBLE 👁️

Use `--headed` AND `--profile` together!

# ✅ CORRECT - Visible browser + persists login/cookies
agent-browser --profile ~/.agent-browser-profile --headed open "https://higgsfield.ai/create/edit"
 
# ❌ WRONG - Headless mode, user cannot see/interact
agent-browser open "https://higgsfield.ai/create/edit"

User MUST see the browser to:

• Monitor edit progress
• Solve CAPTCHA if it appears ⭐
• Verify output quality

KEY: Batch Processing

DO NOT refresh between edits!

1. Enable unlimited toggle ONCE
2. Upload video → Select edit mode → Apply
3. Wait for result
4. Upload new video → Apply
5. Repeat without refreshing

URLs

• Edit/Create: https://higgsfield.ai/create/edit
• Cinema Studio: https://higgsfield.ai/cinema-studio

Tools

agent-browser --profile ~/.agent-browser-profile --headed open "https://higgsfield.ai/create/edit"
agent-browser snapshot
agent-browser screenshot /path/to/output.png
agent-browser click @eXX
agent-browser type @eXX "text"

UNLIMITED Toggle

Location: LAST control on the right, AFTER quantity +/- buttons.

Enable ONCE, reuse many times:

1. Select unlimited model
2. Click switch with aria-checked="false"
3. Verify aria-checked="true"
4. DO NOT refresh - just upload new video and edit

Image Hosting for Upload

tmpfiles.org API:

curl -F "file=@/path/to/video.mp4" https://tmpfiles.org/api/v1/upload

Edit Modes

YourBrand Mascot Character ⚠️

Mascot Image Path:

/path/to/mascot.png

⚠️ ERROR HANDLING:

If user asks to use the mascot, you will see:

Cannot read "mascot.png" (this model does not support image input)

Solution: Use text prompts to describe mascot scenes instead.

Batch Workflow

# 1. Open edit page (visible + persists cookies)
agent-browser --profile ~/.agent-browser-profile --headed open "https://higgsfield.ai/create/edit"
sleep 3
 
# 2. Enable unlimited (ONE TIME)
agent-browser click @unlimited_switch
sleep 1
 
# 3. Edit video 1
# Upload video via drag-drop or file input
agent-browser click @upload_zone
# Select video file...
 
agent-browser click @edit_mode_dropdown
agent-browser click @extend_mode
agent-browser click @apply_button
echo "Edited video 1"
 
# 4. Edit video 2 (NO REFRESH!)
agent-browser click @upload_zone
# Select new video...
 
agent-browser click @edit_mode_dropdown
agent-browser click @loop_mode
agent-browser click @apply_button
echo "Edited video 2"
 
# 5. Continue for more videos...

CAPTCHA HANDLING ⭐ CRITICAL

IMPORTANT: Higgsfield WILL show CAPTCHA after clicking Apply/Edit button.

When CAPTCHA appears, you MUST:

1. Play alert sound:

python3 -c "import os; [os.system('printf \"\\a\" * 3)' for _ in range(1)]" 2>/dev/null || echo -e "\a\a\a"

1. Speak to user:

elia-speak -e "Captcha detected! Please fill the captcha in the browser window, then I'll continue."

1. Wait for user to solve:

sleep 10

1. After captcha solved, RE-CLICK Apply:

agent-browser click @apply_button

Full Captcha Alert Script:

# Check if captcha appeared
if agent-browser snapshot | grep -qi "captcha\|verification\|verify"; then
    echo -e "\a\a\a"  # Beep beep
    elia-speak -e "Captcha detected! Please fill the captcha in the browser window, then I'll continue."
    echo "Waiting for user to solve captcha..."
    sleep 10  # Give user time to solve
    # Re-click Apply after solving
    agent-browser click @apply_button
fi

After Captcha Resolution:

# Wait for page load
sleep 3
 
# Wait for edit to complete
sleep 60
 
# Check Downloads
ls -lath ~/Downloads/hf_*.mp4

Tips

• NO REFRESH between edits
• Always enable unlimited toggle for free models
• Take screenshot after batch complete
• After captcha, RE-CLICK Apply to trigger edit

DOWNLOADING VIDEOS

Where Downloads Go

• Files save to ~/Downloads/ (not configurable)
• Filename pattern: hf_[YYYYMMDD]_[HHMMSS]_[UUID].mp4

Method: Bulk Download from Assets Folder

Steps:

1. Navigate to: https://higgsfield.ai/assets
2. Click on your project folder
3. Select edited videos using checkboxes
4. Click "Download" button
5. Files save to ~/Downloads/

Checking Downloaded Files

# List recent video downloads
ls -lath ~/Downloads/hf_*.mp4 | head -10
 
# Copy to project folder
cp ~/Downloads/hf_*.mp4 /path/to/project/

Tips

• NO REFRESH between edits
• Always enable unlimited toggle for free models
• Take screenshot after batch complete
• Download from Assets page - select checkboxes, click Download
• Files save to ~/Downloads/ automatically with UUID in filename