li-transcript

Extract cleaned video transcripts and archive them with platform metadata.

79|18|Updated Mar 24, 2026
One-click install
npx skills add https://github.com/jiangjiax/li-skills --skill li-transcript
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: li-transcript
Source: https://github.com/jiangjiax/li-skills/tree/main/li-transcript
Command: npx skills add https://github.com/jiangjiax/li-skills --skill li-transcript

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires yt-dlp, ffmpeg, tencentcloud-sdk-python-asr, and includes scripts (resource) components.

What problem does it solve?

li-transcript eliminates the manual effort of transcribing videos and cleaning up speech-to-text errors so you can quickly extract the actual script for benchmarking and analysis.

Core Features & Use Cases

  • Video link → raw transcript: Downloads the video audio (via yt-dlp), converts it to a compatible 16kHz mono format (via ffmpeg), and runs Tencent Cloud ASR to get a transcript.
  • AI proofreading for common ASR mistakes: Cleans obvious recognition errors (homophones/near-sounds, proper nouns, and garbled segments) while preserving spoken wording and meaning.
  • Author identification & benchmark archiving: Detects the author from the transcript (or uses user-provided author info) and saves the result into the matching benchmark blogger directory under the correct time-based folder structure.
  • Platform-aware metadata: Chooses the right YAML section naming based on the URL domain (e.g., 小红书/B站/抖音/YouTube/other).

Quick Start

Ask the skill to extract and proofread the transcript from a given video link, then archive it under the appropriate benchmark blogger folder.

Frequently Asked Questions about li-transcript

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract a video transcript and archive it by competitor for benchmarking?

To extract a video transcript, the skill downloads audio via yt-dlp, converts it to 16kHz mono using ffmpeg, and runs Tencent Cloud ASR. It then cleans the transcript and saves it into a platform-specific YAML markdown file archived by competitor.

How does AI proofreading clean up speech-to-text ASR mistakes in video transcriptions?

AI proofreading cleans speech-to-text ASR mistakes by targeting homophones, near-sounds, proper nouns, and garbled segments. It corrects these recognition errors while preserving the original spoken wording and meaning for accurate content archiving.

Do I need to install yt-dlp and ffmpeg to get transcripts from video links?

Yes, you need yt-dlp and ffmpeg installed to get transcripts from video links. The skill depends on yt-dlp for retrieving video audio and ffmpeg for converting it to the 16kHz mono format required by Tencent Cloud ASR.

Can I use Tencent Cloud ASR to transcribe YouTube or B站 videos?

Yes, you can use Tencent Cloud ASR to transcribe YouTube or B站 videos. The skill supports platform-aware metadata extraction, automatically selecting the correct YAML naming conventions based on the URL domain provided.

What is the best way to archive video transcriptions for competitor benchmarking?

The best way to archive video transcriptions for competitor benchmarking is to save them as markdown files with YAML metadata. This skill detects the author from the transcript and organizes the files into matching benchmark directories using a time-based folder structure.