AI Agent Use Cases
Views —

AI Agent Use Cases

Added today

44All 189

Prompt to Prepare a Shopping Cart from a Recipe Photo

The author reports that Grok Bot receives a recipe photo, organizes the ingredients and quantities, and creates a shopping list by excluding ingredients already available at home.

Grok Bot
Life & Hobbies

1–20 of 189

Marketing & Design

Automating the Removal of Pauses from Recorded Videos

This is a report on how Claude Code was used to automatically select segments to cut based on transcribed content, streamlining the editing of short videos.

Claude Code

How it works

  1. Ran whisper.cpp locally to transcribe the recorded video, then used Claude Code to select which segments to cut based on the transcription.
  2. Used ffmpeg to cut and render the video according to the selected segments.
  3. After rendering, extracted frames from each cut and measured audio levels to check the results. All rules to follow were collected in the repository's HARD-RULES.md.

Results

The author reported that the original 118-second video was reduced to 59.5 seconds. 42% of the original consisted of pauses and retakes.

Notes

Most of the time was spent catching renders that were technically error-free but incorrect, due to issues like vertical video rotation values, order of speed-up options, audio glitches at cut boundaries, and frame rate mismatches. The repository link is only mentioned as being in the original comments.

Source

Reddit Original post ↗

Example prompt · Claude Code

Implement an editing pipeline that reduces long silences and repeated speech segments in recorded videos.

Input: Original video [file path], sentences to keep [criteria], target length [duration], output aspect ratio [width/height]. First, investigate the video's actual display orientation, resolution, frame rate, and audio format.

Use whisper.cpp locally to create a timestamped transcription, and based on that, list candidate segments for deletion along with reasons. Exclude edits that would change the meaning of the speech, and only cut confirmed segments using ffmpeg, saving them as separate files. Preserve both the original and the edit segment list so changes can be reverted.

Check for vertical video rotation info, timing of speed-up application, noise at cut boundaries, and frame rate differences in overlays. For each cut, extract a verification frame and check audio levels to detect cases where commands succeeded but results are incorrect.

Deliverables: Execution script, dependency installation instructions, segment list, edited video, and verification log. First, run a short sample to check display orientation, audio continuity, and final length before processing the entire file.