ytx-cli 0.1.0

Extract YouTube transcripts from the terminal. Pipe-friendly, no API key needed.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89

# ytx

Extract YouTube transcripts from the terminal. Pipe-friendly, no API key needed.

## Why ytx?

- **Stdout by default** -- transcripts go straight to your terminal, ready to pipe anywhere
- **No API key, no setup** -- scrapes what it needs from the video page automatically
- **Built for LLM workflows** -- feed any YouTube video into `llm`, `fabric`, `mods`, or your own tools in one command

## Install

```bash
cargo install ytx-cli
```

Or build from source:

```bash
git clone https://github.com/koguchic/ytx.git
cd ytx && cargo install --path .
```

## Quick Start

```bash
ytx dQw4w9WgXcQ                         # bare video ID
ytx https://youtu.be/dQw4w9WgXcQ        # short URL
ytx https://youtube.com/watch?v=...      # full URL
ytx dQw4w9WgXcQ -t                      # with timestamps
ytx dQw4w9WgXcQ -l ja                   # Japanese captions
```

## Pipe It Anywhere

This is where `ytx` shines. It's a Unix building block -- grab a transcript and send it wherever you want.

```bash
# Search a lecture for a specific concept
ytx https://youtube.com/watch?v=... | grep -i "attention mechanism"

# Summarize a video with an LLM
ytx https://youtube.com/watch?v=... | llm "what are the key takeaways?"

# Extract action items with fabric
ytx https://youtube.com/watch?v=... | fabric -p extract_wisdom

# Get a word count
ytx https://youtube.com/watch?v=... | wc -w

# Save to a file
ytx https://youtube.com/watch?v=... > transcript.txt

# Copy to clipboard (macOS)
ytx https://youtube.com/watch?v=... | pbcopy

# Translate via LLM
ytx https://youtube.com/watch?v=... -l ja | llm "translate this to English"

# Timestamped transcript to file
ytx https://youtube.com/watch?v=... -t -o notes.txt
```

## Options

| Flag | Description | Default |
|------|-------------|---------|
| `-t, --timestamps` | Prepend `[MM:SS]` or `[HH:MM:SS]` timestamps | off |
| `-l, --lang <CODE>` | Caption language code | `en` |
| `-o, --output <FILE>` | Write to file instead of stdout | stdout |

## How It Works

1. Fetches the YouTube video page and extracts the InnerTube API key
2. Calls the InnerTube player API (ANDROID client) to get caption URLs -- no PoToken needed
3. Fetches and parses the caption XML (supports both srv3 and legacy formats)

No API key required. No browser automation. Just HTTP requests.

## Supported URL Formats

- `https://www.youtube.com/watch?v=VIDEO_ID`
- `https://youtu.be/VIDEO_ID`
- `https://www.youtube.com/shorts/VIDEO_ID`
- Bare 11-character video IDs

## License

MIT