By name: summarize description: Summarize or extract text from URLs, local files, YouTube videos, and podcasts using the summarize CLI. Use when the user asks to summarize a link, article, video, or document. Supports multiple output lengths and models. Do NOT use for web search (use web-search) or for content the agent can read directly. metadata: model: remote tools: - exec - Bash triggers: - 요약 - 정리 - summarize - 요약해줘 - 정리해줘 aliases: - 요약
homepage: https://summarize.sh
Summarize
Quick Reference
| Input | Command |
|---|---|
| URL/article | summarize “https://...” --model openai/gpt-4.1-mini |
| Local file | summarize “/path/to/file.pdf” |
| YouTube | summarize “https://youtu.be/...” --youtube auto |
| Extract only | summarize “URL” --extract-only |
Quick start
summarize "https://example.com" --model openai/gpt-4.1-mini
summarize "/path/to/file.pdf" --model anthropic/claude-3-5-sonnet
summarize "https://youtu.be/dQw4w9WgXcQ" --youtube auto
YouTube: summary vs transcript
Best-effort transcript (URLs only):
summarize "https://youtu.be/dQw4w9WgXcQ" --youtube auto --extract-only
If the user asked for a transcript but it’s huge, return a tight summary first, then ask which section/time range to expand.
Model + keys
Set the API key for your chosen provider:
- OpenAI:
OPENAI_API_KEY - Anthropic:
ANTHROPIC_API_KEY
Default model is openai/gpt-4.1-mini if none is set.
Useful flags
--length short|medium|long|xl|xxl|<chars>--max-output-tokens <count>--extract-only(URLs only)--json(machine readable)--youtube auto
Config
Optional config file: ~/.summarize/config.json
{ "model": "openai/gpt-5.2" }
Optional services:
- Additional provider-specific options can be configured in summarize CLI docs.