返回顶部
a

ai-video-summarizer

>

作者: admin | 来源: ClawHub
源自
ClawHub
版本
V 1.0.0
安全检测
已通过
132
下载量
0
收藏
概述
安装方式
版本历史

ai-video-summarizer

# AI Video Summarizer — Turn Any Video into Actionable Summaries The world produces 500 hours of YouTube video every minute. Podcast episodes average 60-90 minutes. Conference talks run 30-45 minutes. Lecture recordings span 50-90 minutes. Webinar replays last 45-120 minutes. The knowledge inside these videos is valuable — but accessing it requires watching at 1x speed, hoping you don't miss the key insight buried at minute 37 of a 60-minute recording. Nobody has time to watch everything. But everyone needs the insights. NemoVideo bridges this gap by watching the video for you. Upload any video — a YouTube URL, a podcast episode, a lecture recording, a meeting replay, a conference talk — and the AI produces: a concise summary (the entire video's value in 200-500 words), timestamped key points (the 5-10 most important insights with exact timestamps), chapter breakdown (every topic change with a description), highlight clips (the 3-5 most valuable segments extracted as standalone videos), full transcript with topic segmentation, and a one-sentence TL;DR that captures the video's core message. Hours of content become minutes of reading — with the ability to jump to any specific moment that matters. ## Use Cases 1. **YouTube Video → Key Takeaways (any length)** — A 25-minute YouTube video about investment strategies. NemoVideo produces: a 300-word summary covering all key strategies, 7 timestamped key points ("At 8:22: Dollar-cost averaging outperforms lump-sum investing 68% of the time in volatile markets"), chapter breakdown with 5 topics, the 3 strongest clips extracted as standalone videos (each 30-90 seconds), and a TL;DR: "Consistent monthly investing in index funds with annual rebalancing outperforms active management for 95% of individual investors." 2. **Podcast Episode → Show Notes (30-120 min)** — A 90-minute interview podcast. NemoVideo generates: summary covering the guest's main arguments, 10 quotable moments with timestamps and speaker labels ("At 34:15, Dr. Chen: 'The real breakthrough wasn't the algorithm — it was realizing we were asking the wrong question'"), topic timeline showing when each subject was discussed, 5 highlight clips for social media promotion, and a full transcript with speaker identification. 3. **Lecture Recording → Study Notes (50-90 min)** — A 75-minute university lecture on organic chemistry. NemoVideo produces: structured notes following the lecture's progression, all definitions extracted with timestamps, key formulas and concepts highlighted, practice problems identified and timestamped, a summary that reads like a textbook chapter overview, and links between concepts ("This reaction mechanism connects to the SN2 reactions covered at 12:30"). 4. **Conference Talk → Executive Brief (20-45 min)** — A 35-minute keynote the CEO couldn't attend. NemoVideo delivers: a 200-word executive summary, 5 actionable takeaways, 3 data points or statistics cited, the speaker's key recommendation, and a 2-minute highlight clip capturing the most important moment. The CEO gets the talk's full value in 3 minutes of reading. 5. **Meeting Recording → Action Items (15-120 min)** — A 60-minute team meeting. NemoVideo extracts: all decisions made (with timestamps and participants), all action items (with assignee, deadline, and context), all open questions flagged for follow-up, a 150-word executive summary, and topic-by-topic breakdown. The meeting becomes an actionable document instead of a forgotten recording. ## How It Works ### Step 1 — Provide Video Upload a video file or paste a URL. NemoVideo accepts any format, any length, any language. ### Step 2 — Choose Summary Outputs Select: text summary, key points, chapters, highlight clips, transcript, action items, or all of them. ### Step 3 — Generate ```bash curl -X POST https://mega-api-prod.nemovideo.ai/api/v1/generate \ -H "Authorization: Bearer $NEMO_TOKEN" \ -H "Content-Type: application/json" \ -d '{ "skill": "ai-video-summarizer", "prompt": "Summarize this 45-minute conference talk about AI in healthcare. Generate: 300-word summary, top 8 key points with timestamps, chapter breakdown, 3 highlight clips (best 60-second segments), full transcript with speaker labels, and a one-sentence TL;DR. Focus on actionable insights for hospital administrators.", "outputs": ["summary", "key-points", "chapters", "highlight-clips", "transcript", "tldr"], "summary_length": "300 words", "key_points_count": 8, "highlight_clips_count": 3, "highlight_clip_duration": "60 sec", "focus": "actionable insights for hospital administrators", "language": "en" }' ``` ### Step 4 — Review and Use Review summary accuracy. Jump to any timestamped moment. Share highlight clips. Distribute the summary to stakeholders. ## Parameters | Parameter | Type | Required | Description | |-----------|------|:--------:|-------------| | `prompt` | string | ✅ | Video source and summary requirements | | `outputs` | array | | ["summary","key-points","chapters","highlight-clips","transcript","tldr","action-items"] | | `summary_length` | string | | "brief" (100w), "standard" (300w), "detailed" (500w) | | `key_points_count` | integer | | Number of key points (default: 5-10 based on length) | | `highlight_clips_count` | integer | | Number of clips to extract (default: 3) | | `highlight_clip_duration` | string | | "30 sec", "60 sec", "90 sec" | | `focus` | string | | Specific angle or audience to prioritize | | `language` | string | | "auto", "en", "es", "de", "fr", "ja", "zh" | | `speaker_labels` | boolean | | Identify speakers (default: true) | | `batch_urls` | array | | Multiple videos for batch summarization | ## Output Example ```json { "job_id": "avs-20260328-001", "status": "completed", "source_duration": "45:18", "language": "en", "outputs": { "tldr": "AI diagnostic tools reduce emergency department wait times by 34% when integrated into existing triage workflows rather than replacing them.", "summary": "Dr. Sarah Chen's keynote argues that AI in healthcare succeeds not by replacing clinical judgment but by augmenting it at specific decision points... [300 words]", "key_points": [ {"timestamp": "3:22", "point": "AI triage tools reduce ED wait times by 34% in the Mount Sinai pilot"}, {"timestamp": "8:45", "point": "Integration failures account for 78% of failed healthcare AI deployments"}, {"timestamp": "15:10", "point": "The 3-phase implementation model: shadow mode → advisory mode → autonomous mode"} ], "chapters": [ {"timestamp": "0:00", "title": "The Promise vs Reality of Healthcare AI", "duration": "6:30"}, {"timestamp": "6:30", "title": "Case Study: Mount Sinai ED Triage", "duration": "8:15"}, {"timestamp": "14:45", "title": "The 3-Phase Implementation Framework", "duration": "10:20"} ], "highlight_clips": [ {"file": "highlight-1.mp4", "timestamp": "8:22-9:24", "topic": "34% wait time reduction statistic"}, {"file": "highlight-2.mp4", "timestamp": "15:10-16:08", "topic": "3-phase framework introduction"}, {"file": "highlight-3.mp4", "timestamp": "38:45-39:50", "topic": "Closing recommendation"} ] } } ``` ## Tips 1. **Key points with timestamps are more useful than full transcripts** — Most people don't read 7,000 words. They want the 5 insights that matter, with the ability to jump to the exact moment for context. 2. **Highlight clips are instant social content** — The 3 best 60-second moments from a 45-minute talk become LinkedIn posts, Twitter clips, or newsletter embeds. One summary operation produces a week of promotional content. 3. **Focus parameter changes everything** — Summarizing the same talk "for investors" vs "for engineers" vs "for customers" produces three different summaries emphasizing different points. The same content, different audiences. 4. **Batch summarization turns a conference into a document** — 20 conference talks → 20 summaries → one combined brief. An executive gets the entire conference's value in 30 minutes of reading. 5. **Action items extraction turns meetings into accountability** — "Sarah will send the proposal by Friday" extracted automatically means nobody can claim they didn't hear the commitment. ## Output Formats | Format | Content | Use Case | |--------|---------|----------| | MD | Formatted summary + key points | Reading / sharing | | JSON | Structured data | API integration | | MP4 | Highlight clips | Social media / promotion | | SRT | Timestamped transcript | Reference / captions | | TXT | Plain text summary | Email / messaging | ## Related Skills - [ai-video-script-generator](/skills/ai-video-script-generator) — Write video scripts - [video-hook-maker](/skills/video-hook-maker) — Generate hooks - [video-color-correction-ai](/skills/video-color-correction-ai) — Color correction

标签

skill ai

通过对话安装

该技能支持在以下平台通过对话安装:

OpenClaw WorkBuddy QClaw Kimi Claude

方式一:安装 SkillHub 和技能

帮我安装 SkillHub 和 ai-video-summarizer-1775964002 技能

方式二:设置 SkillHub 为优先技能安装源

设置 SkillHub 为我的优先技能安装源,然后帮我安装 ai-video-summarizer-1775964002 技能

通过命令行安装

skillhub install ai-video-summarizer-1775964002

下载 Zip 包

⬇ 下载 ai-video-summarizer v1.0.0

文件大小: 4.6 KB | 发布时间: 2026-4-13 09:17

v1.0.0 最新 2026-4-13 09:17
AI Video Summarizer 1.0.0 — Initial Release

- Summarize any video into key points, chapter breakdowns, highlight clips, and concise text summaries.
- Produces timestamped insights, chapter markers with descriptions, 3–5 top highlight clips, full transcript with topic segmentation, and a one-sentence TL;DR.
- Supports a wide range of use cases: YouTube, podcasts, lectures, conference talks, and meeting recordings.
- Flexible output options: choose which summaries you want, set summary length, highlight clip duration, key points count, and focus.
- Batch summarization and speaker identification supported for efficient analysis of multiple videos.

Archiver·手机版·闲社网·闲社论坛·羊毛社区· 多链控股集团有限公司 · 苏ICP备2025199260号-1

Powered by Discuz! X5.0   © 2024-2025 闲社网·线报更新论坛·羊毛分享社区·http://xianshe.com

p2p_official_large
返回顶部