Video Deconstruction Tool

Paste a Douyin, Kuaishou, or Xiaohongshu video link. Get the transcript, hook analysis, structure breakdown, and pacing report — all in one step.

Try It Free →
⚡

Paste a link, get the full breakdown

You found a video that clearly works. Instead of watching it ten times and taking notes by hand, paste the link into ShenBi AI.

The system does three things automatically:

🎯

Supports Douyin, Kuaishou, and Xiaohongshu

Paste share links from any of the three major Chinese short-video platforms. The system handles video download, audio extraction, and speech-to-text transcription automatically.

Each platform has different pacing norms and expression styles. The analysis accounts for these differences — a Kuaishou video is evaluated against Kuaishou standards, not Douyin benchmarks.

📊

What you get from a deconstruction

After processing, you receive:

🔗

From analysis to new content

Deconstruction is step one in a four-step workflow:

Transcribe → Deconstruct → AI Review → Generate Scripts

The analysis feeds directly into script generation. Instead of asking AI to write from nothing, you give it a real reference foundation — the structural logic of a video that already worked.

Generate scripts for Douyin, Kuaishou, or Xiaohongshu based on the deconstruction results.

📊

Real-world examples

Three example workflows showing input → output:

📥 Example 1: A Douyin knowledge video (topic: "3 ways to fall asleep in 30 seconds", 45-second runtime)

📤 Output includes:

  • Full transcript with 0.5-second timestamp precision
  • 9-cell visual summary (open / middle / close keyframes)
  • Hook classification: number-impact + pain-point resonance
  • Pacing map: 0-3s hook / 3-15s problem / 15-35s solution / 35s+ CTA
  • Reusable expression patterns (with reuse-as-method, not copy-as-content boundary)

📥 Example 2: A Xiaohongshu lifestyle reference video (topic: weekend cafe visit)

📤 Output includes:

  • Full script + scene-based expression nodes
  • Recommendation mechanic analysis (mood-setup → product-intro → experience-detail)
  • First-person narrative pacing markers
  • Reusable "authenticity" patterns + risk flags for verbatim copying

📥 Example 3: A Kuaishou daily-life video (topic: rural countryside daily)

📤 Output includes:

  • Dialect-aware full transcript
  • Authenticity expression mechanics (trust-building / scene-narrative)
  • Pacing characteristics: slower opening + real-detail driven
  • Cross-platform migration notes (slow pacing risks low completion on Douyin)

Note: examples are synthetic scenarios; no real account or video copyright data referenced.

Video Deconstruction Tool FAQ

Which platform video links are supported?

Douyin, Kuaishou, and Xiaohongshu share links. Paste the link from the app or browser directly.

What languages can it transcribe?

The system automatically detects the spoken language. It supports 30 languages including Chinese, English, Japanese, Korean, French, Spanish, German, Arabic, Thai, Vietnamese, Indonesian, Russian, and more. It also recognizes 22 Chinese regional dialects.

How fast is deconstruction?

Transcription and analysis typically complete within seconds for standard-length short videos. Longer videos may take slightly more time.

Is this just a transcript tool?

No. Transcription is one part of it. The tool also analyzes hook type, structure, pacing, emotional arc, and persuasion method. It gives you a complete structural breakdown, not just text.

Can I generate scripts from the analysis?

Yes. The deconstruction results feed directly into ShenBi AI script generation. You can generate platform-adapted scripts based on the structural logic extracted from the reference video.

Try It Free 🚀