Skip to main content

​
Basic usage

notewise process "https://youtube.com/watch?v=VIDEO_ID"
Notes are written to ./output/<sanitized video title>.md by default.

​
Accepted URL formats

# Standard watch URL
notewise process "https://www.youtube.com/watch?v=dQw4w9WgXcQ"

# Short URL
notewise process "https://youtu.be/dQw4w9WgXcQ"

# Shorts URL
notewise process "https://www.youtube.com/shorts/dQw4w9WgXcQ"

# Embed URL
notewise process "https://www.youtube.com/embed/dQw4w9WgXcQ"

# Bare 11-character video ID
notewise process dQw4w9WgXcQ

​
Common options

notewise process "URL" --output ~/my-notes
notewise process "URL" -o ~/my-notes

notewise process "URL" --model gpt-4o
notewise process "URL" --model anthropic/claude-3-5-sonnet-20241022
notewise process "URL" -m groq/llama3-70b-8192

notewise process "URL" --temperature 0.3   # deterministic
notewise process "URL" --temperature 0.9   # creative
notewise process "URL" -t 0.5

notewise process "URL" --quiz
Produces <title>_quiz.md — 10–15 multiple-choice questions with answers and explanations.
notewise process "URL" --export-transcript txt    # plain text
notewise process "URL" --export-transcript json   # JSON with timestamps
See Output Format for the exact file structure.
notewise process "URL" --language hi        # Hindi first, English fallback
notewise process "URL" -l fr -l en          # French, then English
The flag is repeatable — notewise tries each language in order and falls back to the next if unavailable.
By default, videos already in cache with existing output are skipped. Use --force to override:
notewise process "URL" --force
notewise process "URL" -F

notewise process "URL" --no-ui
Disables the Rich live dashboard and prints plain progress lines to stdout. Suitable for cron jobs, CI pipelines, and log aggregators.

​
Output files

output/
├── Video Title.md
├── Video Title_quiz.md              # only with --quiz
├── Video Title_transcript.txt       # only with --export-transcript txt
└── Video Title_transcript.json      # only with --export-transcript json
Chapter-aware videos (longer than 1 hour with chapter data) use a directory layout instead. See Output Format.

​
Inspect without processing

notewise info "https://youtube.com/watch?v=VIDEO_ID"
Fetches title, duration, chapter count, and available transcript languages — without generating any notes.