ffmpeg-ai 0.1.8

AI-powered FFmpeg command generator

FFmpeg AI Assistant

A command-line utility that acts as an intelligent FFmpeg assistant for developers. This tool runs entirely offline and uses a local LLM via Ollama to provide context-aware FFmpeg commands, code snippets, and explanations.

---

Features

• Generate FFmpeg commands from natural language queries
• Get Python, Bash, or Node.js code wrappers for FFmpeg commands
• Detailed explanations of command parameters
• 100% offline operation with local LLM and vector database
• Smart caching for repeated queries
• Easy customization and model switching

---

Installation

Recommended: PyPI Installation

Install the tool from PyPI using:

pip install ffmpeg-ai

---

Documentation Setup After Installation

After installing, you must initialize the documentation database:

ffmpeg-ai setup

Command-Line Interface

Once installed, you can interact with the tool via the command line. Use the following command to get help:

ffmpeg-ai --help

This will display the available options and commands:

Usage: ffmpeg-ai --help

AI-powered FFmpeg command generator

Options:
  --help  Show this message and exit.

Commands:
  query        Ask a natural language question about FFmpeg.
  setup        Set up the FFmpeg documentation database.
  clear-cache  Clear the query cache.

---

Documentation Setup

Running ffmpeg-ai setup will:

• Download and parse FFmpeg's official documentation
• Create vector embeddings for faster and more accurate querying
• Store the embeddings locally for offline use

---

Real Usage Example

Here's an example of using the tool in your project:

# In your project directory, run:
ffmpeg-ai query "convert .mov to .mp4 using H.264 codec" --code

Result:

FFmpeg Command:

ffmpeg -i input.mov -vcodec libx264 output.mp4

Generated Python Script:

import os
import subprocess

def convert_mov_to_mp4(input_file, output_file):
    command = f'ffmpeg -i {input_file} -vcodec libx264 {output_file}'
    subprocess.run(command, shell=True, check=True)

# Usage example:
input_file = 'input.mov'
output_file = 'output.mp4'
convert_mov_to_mp4(input_file, output_file)

---

How It Works

1. Document Retrieval: Fetches the most relevant chunks from embedded FFmpeg documentation using sentence-transformers and ChromaDB.
2. Query Augmentation: Combines your query with retrieved context before sending it to the local LLM.
3. LLM Generation: The local model (e.g., Mistral) generates the command, code, and explanations.
4. Result Formatting: Outputs are formatted cleanly and cached for faster future retrieval.

---

Requirements

• Python 3.10+
• FFmpeg installed system-wide
• Ollama installed and running
• Local machine with ~8GB RAM or more recommended

---

Limitations

• Responses depend on the quality of the local LLM model.
• Currently limited to embedded and user-added FFmpeg documentation.
• Hardware-intensive for optimal performance.

---

License

MIT License