tube-summary

**Judgment** `tube-summary` is a solid, narrowly focused utility for turning YouTube videos into structured text summaries using subtitles. It’s well-suited for agents that need to turn long-form video content into digestible text (topics, timestamps, quotes) but depends heavily on the availability and quality of English subtitles and on YouTube’s current behavior. **What it does well** - Provides an end-to-end flow: search → user selects a result → download English subtitles → process and summarize. - Produces structured outputs (key topics, 2–3 paragraph summary, key timestamps, key quotes), which are useful for downstream reasoning or note-taking. - Efficient: uses yt-dlp to fetch subtitles only, avoiding full video downloads. - Implementation is straightforward Python with clear scripts, making it relatively easy to debug or extend. **Key limitations / risks** - **Subtitle dependence**: If a video has no English subtitles (or low-quality auto-generated ones), summaries will be missing or inaccurate. - **YouTube fragility**: Relies on YouTube search (or scraping) and yt-dlp; API or site changes, rate limits, or regional restrictions can break search or subtitle download. - **Language limitation**: Explicitly targets English subtitles only; not suitable for non-English content. - **Content safety**: It will summarize whatever content the user selects, including potentially harmful, misleading, or explicit material, with no built-in filtering or moderation. - **TOS / legal considerations**: Use of yt-dlp and scraping-based search may raise terms-of-service or legal concerns depending on environment and policies. **Recommended scenarios** - Agents assisting with **study, research, or technical learning** that need concise overviews of long YouTube talks or tutorials. - **Productivity / note-taking agents** that convert videos into structured notes with timestamps and quotes for later review. - **Content triage**: summarizing several candidate videos on a topic so a user can decide which ones to watch fully. **Less suitable scenarios** - Environments with strict compliance or legal requirements around YouTube access, scraping, or yt-dlp usage. - Use cases requiring reliable support for **non-English** content or videos without any subtitles. - High-stakes domains where **content safety and factual verification** must be enforced by the tool itself, rather than by the surrounding agent logic.