Video To GIF#

Quick workflow#

1. Identify the input video path or ask the user to upload/provide it if missing.
2. Choose sensible defaults unless the user specifies otherwise:
   • fps=12
   • keep the original video resolution by default; only resize when the user asks for compression/smaller file size or a target dimension
   • trim duration to ≤10 seconds if the user asks for a generic “动图” but the video is long
   • use palette generation for GIF quality
3. Run scripts/video_to_gif.py from this skill directory.
4. Verify the output exists and report path/size. Attach with MEDIA:<path> if delivering in chat.

Script usage#

bash

python3 /root/.openclaw/workspace/skills/video-to-gif/scripts/video_to_gif.py \
  input.mp4 output.gif \
  --start 00:00:02 --duration 4 \
  --fps 12 --width 480

Useful options:

• --start: clip start time, e.g. 3.5 or 00:00:03.500
• --duration: clip length in seconds or timestamp format
• --end: alternative to duration
• --fps: lower values reduce size; 10–15 is usually good
• --width: scale output width while preserving aspect ratio
• --height: scale output height while preserving aspect ratio
• --output-format gif|webp: WebP is usually much smaller than GIF
• --loop: 0 means loop forever
• --max-colors: GIF palette colors, default 256; lower reduces size

Quality and size guidance#

• Default to original resolution for clarity.
• For Feishu/chat stickers or size-sensitive sharing, ask/choose a smaller width such as --width 360 --fps 10.
• If GIF is too large, reduce in this order: duration, width, fps, colors.
• If the user only needs an animated image and not strict GIF, recommend WebP because it is smaller and smoother.

Dependencies#

The script requires ffmpeg and ffprobe on PATH. If missing, tell the user the dependency is absent and do not attempt risky system installation without confirmation.