content-summarizer 1.2.1

A command-line tool to summarize YouTube videos using AI.

Content Summarizer

Summarize YouTube videos in seconds using AI (Google Gemini or Ollama).

Content-Summarizer is a CLI tool that downloads YouTube audio, transcribes it locally using Faster-Whisper (or uses captions if available), and generates dense, structured summaries using LLMs.

---

Installation

Prerequisites

• Python 3.11+
• FFmpeg (Required for audio processing):
  • Windows: winget install ffmpeg
  • Linux: sudo apt install ffmpeg
  • macOS: brew install ffmpeg

Installing (Recommended)

Use uv or pipx to install in an isolated environment:

# Via UV
uv tool install content-summarizer

# Via Pipx
pipx install content-summarizer

---

First Steps

Before running your first summary, you must configure your default AI provider. You only need to do this once.

Option A: Using Google Gemini

1. Get your free key at Google AI Studio.
2. Run:

content-summarizer config --provider gemini --gemini-key "YOUR_API_KEY_HERE" --gemini-model "3-flash"

Option B: Using Ollama

1. Ensure you have Ollama installed and running (ollama serve).
2. Run (Example using Mistral, ensure you ran ollama pull mistral first):

content-summarizer config --provider ollama --ollama-model "mistral" --ollama-ctx 16384

---

Usage

The application has two main commands: summarize (the main action) and config (to set persistent defaults).

1. The summarize Command

Fetches, processes, and summarizes a video.

Syntax:

content-summarizer summarize "YOUTUBE_URL" [OPTIONS]

Common Options:

• Core:

  • -o, --output-path <path>: Directory to save the final .md file.
  • -c, --keep-cache: Keep temporary files (audio/transcription) after finishing.
  • --no-terminal: Do not print the summary to the console.
  • -q: Decrease verbosity (warnings only / -qq for silent).

• Audio & Transcription:

  • -s, --speed-factor <float>: Accelerate audio (e.g., 2.0 for 2x speed). Saves transcription time.
  • -w, --whisper-model <model>: Model size (tiny, base, small, medium, large-v2).
  • --device <device>: Hardware (cuda, cpu, mps).

• AI Providers:

  • -p, --provider <name>: gemini or ollama.
  • -g, --gemini-model <name>: E.g., 3-flash, 3-pro.
  • --ollama-model <name>: E.g., mistral, llama3.

---

2. The config Command

Sets default values so you don't have to type flags every time.

Syntax:

content-summarizer config [OPTIONS]

Examples:

# Set default download path
content-summarizer config -o "path/to/summaries"

# Set default audio speed to 1.5x
content-summarizer config -s 1.5

# Switch default provider to Ollama
content-summarizer config -p ollama --ollama-model mistral

---

Application Defaults

If you do not provide specific flags or configuration, the application uses the following internal default values:

Parameter	Default Value	Description
Provider	gemini	The AI service used for summarization.
Gemini Model	3-flash	The specific Google Gemini model version.
Whisper Model	base	The local model size used for transcription.
Speed Factor	1.25	Audio acceleration (1.25x speed).
Beam Size	5	Beam search size for transcription decoding.
Device	auto	cuda (GPU), mps (Mac), or cpu.
Compute Type	auto	Quantization (e.g., int8, float16).
Ollama URL	http://localhost:11434/v1	Local address for Ollama server.
Ollama Context	16384	Context window size for local LLMs.

---

Configuration Priority

The application resolves settings in the following order (highest to lowest priority):

1. Command-line Flags: (e.g., passing -s 2.0 overwrites everything).
2. Environment Variables: Loaded from .env.
3. User Configuration: Values set via the config command.
4. Application Defaults: Internal hardcoded values.

---

Important Notes

• Remote API: The remote transcription API support (--api, --api-url) is DEPRECATED and will be removed in version 2.0.0. Please migrate to local processing (Whisper).

---

License

This project is licensed under the Apache 2.0 License - see the LICENSE file for details.