Extract decisions, todos, knowledge, preferences, and risks from AI chat sessions into structured memory. Outputs Markdown, JSON, or Obsidian notes.
文档
chatpipe-export
试用Export any AI chat session to Markdown, JSON, or plain text using ChatPipe v1.1. Supports 7 input formats including OpenClaw sessions, Hermes agent logs, ChatGPT exports, and Chatbox exports. Use when the user asks to export/backup their chat history, convert conversations to Obsidian/Notion, or migrate between AI platforms. Free skill wraps the open-source ChatPipe engine. Triggers: "导出聊天记录", "export chat", "save conversation", "backup my chats", "导出到Obsidian", "导出到Notion", "convert to markdown", "chat export".
它能做什么
Export any AI chat session to Markdown, JSON, or plain text using ChatPipe v1.1. Supports 7 input formats including OpenClaw sessions, Hermes agent logs, ChatGPT exports, and Chatbox exports. Use when the user asks to export/backup their chat history, convert conversations to Obsidian/Notion, or migrate between AI platforms. Free skill wraps the open-source ChatPipe engine. Triggers: "导出聊天记录", "export chat", "save conversation", "backup my chats", "导出到Obsidian", "导出到Notion", "convert to markdown", "chat export".
技能文档
ChatPipe Export — One-Click AI Chat Export
Export any AI chat session to clean Markdown, JSON, or plain text. No manual file handling. Just say "export this conversation".
Supported Formats
Input (auto-detected)
| Format | Source |
|---|---|
openclaw-json | Current / past OpenClaw sessions |
hermes-json | Hermes agent logs |
chatgpt-json | ChatGPT data export ZIP |
chatbox-html | Chatbox exported HTML |
chatbox-json | Chatbox conversations.json |
markdown | Role-labeled .md files |
plain-text | Role-labeled .txt files |
Output
| Format | Use case |
|---|---|
markdown | Obsidian, Notion, GitHub, VS Code |
chatgpt-json | Import into ChatGPT tools, API processing |
chatbox-json | Import back into Chatbox |
plain-text | grep, shell scripts, minimal storage |
Quick Export
To export the current session as Markdown:
python3 scripts/export_session.py --markdown
To export a past session by session key:
python3 scripts/export_session.py agent:main:main --format markdown
To export a Hermes session:
python3 scripts/export_session.py agent:hermes:main --format markdown
How It Works
- Fetch the session transcript via OpenClaw API or file system
- Pass it through ChatPipe (auto-detects format)
- Save the output to the workspace
The helper script scripts/export_session.py wraps both the data extraction
and ChatPipe conversion into a single command.
When to Use
| User says | Do this |
|---|---|
| "导出这段聊天" | Export current session as Markdown |
| "保存到Obsidian" | Export to Markdown, tell user where the file is |
| "backup my AI chats" | Export all recent sessions as Markdown batch |
| "convert to JSON" | Export current session as chatgpt-json |
| "导出和Akira的对话" | Find Hermes session, export as Markdown |
Output Location
Exported files are saved to the workspace root with descriptive names:
session-{key}-{date}.md
Never delete or move exported files unless the user explicitly asks.
相关技能
Export a Codex session JSONL from ~/.codex/sessions into a clean Markdown transcript in ~/Documents/Exports. Use when the user wants to export, save, or conv...
当用户要导出并分析与 Claude.ai 网页版的全部聊天历史(用于回顾、检索、复盘、生成报告、做时间线、找具体某次对话)时,加载本 skill。典型触发:用户提到 "claude 聊天记录"、"claude conversations.json"、"导出 claude 对话"、"分析我和 claude 的对话"...
Export a Codex CLI or Codex Desktop (App) session to Markdown, HTML, or Obsidian notes. Use when the user asks to export, save, share, or review a past chat, session, or transcript. Supports --brief, --redact, --since/--until, --grep, --append (incremental), --interactive (message picker), --sessions (merge), --watch (auto-incremental), and --list. Works with both Codex CLI and Codex Desktop app sessions.
当用户要导出并分析与 Gemini(Google)或 ChatGPT(OpenAI)网页版的全部聊天历史时,加载本 skill。典型触发:用户提到 "导出 gemini 对话"、"gemini 聊天记录"、"chatgpt 数据导出"、"OpenAI conversations.json"、"我和 gemini/...
Say "Help me install this skill from https://clawhub.ai/shuminliu026/chat-summary-card" to your local agent, then turn conversations into polished visual sum...