Neura Market
prompt logo

prompt

Free

Agentic video editing via ffmpeg prompts

FreeFree tier
Type
Open Source
Visit Website GitHub (40 stars)

About prompt

A detailed system prompt that defines an 'Agentic Video Editing Engineer' — an AI agent designed to edit video programmatically by reasoning over transcripts, waveforms, and frames, without a traditional timeline interface. The agent uses ffmpeg and Python's PIL to perform operations such as cutting based on speech boundaries, adding overlays, subtitles, transitions, and ensuring production-correct outputs (e.g., lossless segment concatenation, audio fades at cut edges, subtitle positioning). It follows a workflow: inventory, pre-scan, converse, propose, confirm, execute, self-evaluate, iterate, and persist. The prompt emphasizes audio-first editing, artistic freedom within technical constraints, and self-verification of output quality.

Key Features

Audio-first editing: cut candidates come from speech boundaries and silence gaps
Uses ffmpeg and Python PIL for programmatic video manipulation
Lossless segment concatenation with -c copy, avoiding double encoding
Automatic 30ms audio fades at every cut boundary to prevent pops
Subtitle (SRT) alignment using output-timeline offsets
Overlays shifted with setpts to match window start
Word boundary snapping: never cuts inside a word
Padding (30–200ms) to absorb transcript timestamp drift
Artistic freedom: supports split-screen, PiP, lower-thirds, speed ramps, freeze frames, L-cuts, J-cuts, match cuts
Self-verification: agent checks output before presenting to user

Pros & Cons

Pros
  • Detailed production rules ensure high-quality, professional output
  • Audio-first approach reduces unnecessary visual processing and improves edit decisions
  • Artistic freedom allows creative edits without predefined templates
  • Self-verification step prevents shipping flawed output
  • Open source and free to use (requires ffmpeg and Python)
  • Flexible: can handle a wide range of video types and edits
Cons
  • Requires technical knowledge of ffmpeg and command-line tools
  • Not a graphical user interface; all edits are scripted via prompt
  • May be overkill for simple trims or casual editing needs
  • Dependence on LLM reasoning may introduce latency or variability in edits
  • Setup requires Python environment and ffmpeg installation

Best For

Automated video editing based on transcripts and audio analysisContent creation for podcasts, interviews, and tutorialsPost-production workflow where precise cuts on speech are requiredBatch processing of video with consistent production standardsExperimentation with advanced editing techniques (split screen, overlays, etc.)

FAQ

What is the Agentic Video Editing Engineer prompt?
It is a system prompt that turns an LLM into a video editing agent capable of reasoning over transcripts and frames, then executing edits using ffmpeg and Python PIL.
What tools does the agent use?
The agent primarily uses ffmpeg for video/audio processing and PIL for image-based operations (e.g., overlays). It generates structured EDLs and SRT subtitles.
Is this prompt free to use?
Yes, the prompt is part of the open-source awesome-prompts repository on GitHub and is available at no cost.