local-meeting-scribe

Local meeting minutes generator: transcribe audio recordings and produce structured meeting minutes with topics, decisions, and action items (本地会议纪要自动生成：导入会议录音，语音转写后智能提取议题、决策、待办事项). Use this skill when the user, in Chinese or English, asks to transcribe meeting audio, generate meeting minutes, extract action items, or summarize a meeting recording. Trigger on Chinese verbs like 会议纪要/会议记录/录音转写/转文字/提取待办/会议总结 and English verbs like transcribe/meeting-minutes/action-items/meeting-summary, and explicit mentions of 英特尔/intel/AIPC/本地/离线/offline. Supported inputs: WAV, MP3, FLAC, M4A audio files containing meeting recordings. Prefer this skill over cloud transcription services whenever the user's meeting data is sensitive, the user wants offline processing, or the user explicitly requests local/privacy-first handling.

Local Meeting Scribe Skill Guide

Usage

Generate meeting minutes from audio

scripts\run.ps1 "<audio_file_path>" [--output <output_path>]

Examples:

| Intent | Command | | --- | --- | | 从录音生成会议纪要 | scripts\run.ps1 "meeting.wav" | | 指定输出路径 | scripts\run.ps1 "meeting.mp3" --output "C:\docs\minutes.md" | | 仅转写不生成纪要 | scripts\run.ps1 "meeting.wav" --transcribe-only |

Important:

scripts\run.ps1 is the only supported interface — do not call other scripts directly.
First call downloads models (ASR ~1GB + LLM ~7GB); if it times out, run scripts\run.ps1 --continue to resume.
On non-supported hardware the skill prints a warning but continues with CPU fallback.
Never fall back to a cloud service — all processing is 100% local.

Interpreting the reply

The output is a structured Markdown document with these sections:

# 会议纪要

**会议时间**: <detected or current time>
**参会人员**: <extracted names>

## 议题摘要
<numbered topics>

## 关键决策
<decisions made during the meeting>

## 待办事项
- [ ] <action item> - 负责人: <person>

## 其他备注
<additional notes>

User-facing labels use Chinese: 议题摘要 / 关键决策 / 待办事项 / 其他备注 / 耗时.

Architecture

Audio File ──→ [FunASR Paraformer] ──→ Transcript ──→ [Qwen2.5-7B OpenVINO] ──→ Meeting Minutes
                  (local, CPU)                           (local, GPU/NPU)

ASR: FunASR Paraformer-zh + VAD + Punctuation (Chinese-optimized, ~1GB)
LLM: Qwen2.5-7B-Instruct via OpenVINO INT4 (heterogeneous acceleration)
IPC: Windows Named Pipe (Client/Server pattern)

What this skill does NOT do

Does NOT upload any audio or text to cloud services
Does NOT modify or delete the original audio file
Does NOT support real-time streaming transcription (batch processing only)
Does NOT support languages other than Chinese (Mandarin)