← 返回 Skills 市场

Voice Note Transcriber

Name: Voice Note Transcriber
Author: terrycarter1985

作者 terrycarter1985 · GitHub ↗ · v1.1.0 · MIT-0

cross-platform ✓ 安全检测通过

总下载

当前安装

版本数

在 OpenClaw 中安装

/install voice-note-transcriber

功能描述

Fetch voice note emails via IMAP, transcribe attachments with OpenAI Whisper, and save transcripts to an Obsidian vault's fleeting notes folder.

使用说明 (SKILL.md)

Voice Note Transcriber → Obsidian

Fetch voice-note emails from IMAP, transcribe audio attachments with OpenAI Whisper, and save the result as a fleeting note in your Obsidian vault.

Quick start

python3 {skillDir}/scripts/transcribe_voice_notes.py

Configuration

Set environment variables (or let the skill read from OpenClaw config):

Variable	Required	Description
`EMAIL_ADDRESS`	✅	IMAP login address
`EMAIL_PASSWORD`	✅	App-specific password
`OPENAI_API_KEY`	✅	OpenAI API key
`OBSIDIAN_VAULT_PATH`	✅	Absolute path to Obsidian vault root
`EMAIL_PROVIDER`		Provider key: `gmail` (default), `outlook`, `163.com`, `126.com`
`FLEETING_FOLDER`		Vault subfolder for fleeting notes (default: `0-Inbox`)
`VOICE_NOTE_KEYWORD`		Subject filter keyword (default: `Voice Note`)
`MARK_EMAIL_READ`		`true` / `false` (default: `true`)

OpenClaw config example

{
  skills: {
    "voice-note-transcriber": {
      obsidianVaultPath: "/home/user/Documents/MyVault",
      fleetingFolder: "0-Inbox",
      emailProvider: "gmail",
      subjectKeyword: "Voice Note",
      markEmailRead: true,
    },
  },
}

What it does

Connects to IMAP and fetches unread emails whose subject contains the configured keyword.
Downloads audio attachments (.mp3, .wav, .m4a, .webm, .ogg, .flac) to a temp directory.
Transcribes each attachment via the OpenAI Whisper API (whisper-1).
Writes a Markdown note with YAML frontmatter to the vault's fleeting folder.
Optionally marks the source email as read.

Output note format

---
tags:
  - type/transcript
  - source/voice-note
date: 2025-05-12
email_subject: "Voice Note — weekly recap"
email_from: [email protected]
---

# Voice Note Transcript — 2025-05-12

## Transcript

(transcribed text here)

## Source

- **Subject:** Voice Note — weekly recap
- **From:** [email protected]
- **Date:** Mon, 12 May 2025 08:00:00 +0800

Troubleshooting

Symptom	Fix
"Missing OPENAI_API_KEY"	Export `OPENAI_API_KEY` or set it in OpenClaw config
Notes not appearing in Obsidian	Check `OBSIDIAN_VAULT_PATH` is the vault root (not a subfolder) and `FLEETING_FOLDER` exists inside it
No emails found	Verify `EMAIL_PROVIDER`, credentials, and `VOICE_NOTE_KEYWORD` match your mailbox
Attachment not transcribed	Ensure the attachment has a recognised audio extension

Publishing to ClawHub

clawhub publish ~/.openclaw/skills/voice-note-transcriber \
  --slug voice-note-transcriber \
  --name "Voice Note Transcriber" \
  --version 1.0.0 \
  --changelog "Initial release"

安全使用建议

Before installing, make sure you are comfortable giving the skill an email app password and OpenAI key, sending matching voice-note audio to OpenAI, and writing generated transcripts into your Obsidian vault. Use a narrow subject keyword or dedicated mailbox, consider disabling MARK_EMAIL_READ for the first run, and install any missing dependencies from trusted sources.

功能分析

Type: OpenClaw Skill Name: voice-note-transcriber Version: 1.1.0 The skill performs its stated function of fetching voice-note emails via IMAP, transcribing them using the OpenAI Whisper API, and saving the output to an Obsidian vault. The script `scripts/transcribe_voice_notes.py` uses standard libraries like `imbox` and `subprocess` (to call `curl` for the API) in a manner consistent with its documentation, with no evidence of credential exfiltration or malicious command execution.

能力标签

requires-sensitive-credentials

能力评估

✓ Purpose & Capability

The described workflow and code are coherent: unread IMAP messages are filtered by subject, audio attachments are transcribed with OpenAI Whisper, and Markdown notes are written to an Obsidian folder.

ℹ Instruction Scope

The quick-start command runs a batch process over all unread messages matching the keyword, with no per-attachment confirmation or size/count limits, and email read marking defaults to enabled.

ℹ Install Mechanism

There is no install spec, while the script relies on the external Python package imbox and the curl binary in addition to python3.

ℹ Credentials

Mailbox credentials, an OpenAI API key, and an Obsidian vault path are proportionate to the purpose but are sensitive and should be scoped carefully.

ℹ Persistence & Privilege

The skill writes persistent notes into the configured vault and can mutate mailbox read state, but there is no evidence of hidden background persistence or self-start behavior.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install voice-note-transcriber
安装完成后，直接呼叫该 Skill 的名称或使用 /voice-note-transcriber 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v1.1.0

Fix: Obsidian fleeting path now reads OBSIDIAN_VAULT_PATH+FLEETING_FOLDER; real attachment download; proper Whisper API response parsing; configurable via env/config

v1.0.0

Initial release: Whisper transcription → Obsidian daily note

元数据

Slug voice-note-transcriber

版本 1.1.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 2

常见问题