wechat-to-notebooklm

Imported from zstmfhy/wechat-to-notebooklm (SKILL.md). Install upstream with npx skills add zstmfhy/wechat-to-notebooklm. Copyright stays with the author.

WeChat to NotebookLM

Automatically sync WeChat Official Account articles to Google NotebookLM for AI-powered analysis, summarization, and content generation.

What This Does

This skill automates the entire workflow of getting a WeChat article into NotebookLM:

1. Fetches the article content from the URL
2. Converts to clean Markdown format
3. Creates a NotebookLM notebook (optionally with custom title)
4. Uploads the article as a source
5. Returns the notebook ID for further interaction

When to Use

Use this skill when you:

• Have a WeChat Official Account article URL (mp.weixin.qq.com)
• Want to save the article to NotebookLM for analysis
• Want to create a podcast/video from the article
• Want to chat with the article content using AI
• Want to summarize or extract insights from the article

Example triggers:

• "Sync this WeChat article to NotebookLM"
• "Add this mp.weixin.qq.com link to my notebook"
• "Create a notebook from this WeChat article"
• "Save this article to NotebookLM"

Prerequisites

Before using this skill, ensure NotebookLM CLI is authenticated:

notebooklm login          # Opens browser for Google OAuth
notebooklm status         # Verify authentication

Workflow

Complete Sync Process

Step 1: Fetch Article Content

Use the web reader MCP tool to fetch the WeChat article:

mcp__web_reader__webReader
URL: <WeChat article URL>
return_format: markdown
retain_images: false (optional, saves bandwidth)

This returns the article content in Markdown format.

Step 2: Save Content to File

Save the fetched content to a temporary Markdown file:

# Extract title from content or use default
# Save to /tmp/<sanitized_title>.md

The file should be saved with a descriptive name based on the article title.

Step 3: Create NotebookLM Notebook

Create a new notebook with the article title:

notebooklm create "<article_title>" --json

Parse the JSON response to get the notebook ID:

{"notebook": {"id": "abc-123-def", "title": "..."}}

Step 4: Upload Article to Notebook

Add the Markdown file as a source to the notebook:

notebooklm source add /tmp/<article_title>.md --notebook <notebook_id> --json

Parse the response to get the source ID:

{"source": {"id": "source-xyz-789", "title": "...", "type": "text"}}

Step 5: Confirm Success

Report to the user:

• Notebook title and ID
• Source file name
• Notebook ID for further use
• Suggested next steps (ask questions, generate podcast, etc.)

Progress Updates

Provide brief, clear status updates:

✅ Fetching article from WeChat...
✅ Converting to Markdown...
✅ Creating notebook "Article Title"...
✅ Uploading to NotebookLM...
✅ Done! Notebook ID: abc-123-def

Output Summary

When complete, provide:

Successfully synced WeChat article to NotebookLM!

📓 Notebook: [Article Title]
   ID: abc-123-def

📄 Source: article_title.md
   ID: source-xyz-789

💡 Next steps:
   - Use: notebooklm use abc-123-def
   - Ask: notebooklm ask "Summarize this article"
   - Generate: notebooklm generate audio "Create a podcast"

Error Handling

Common Issues

1. Article URL is invalid or inaccessible

• Error: Failed to fetch content
• Solution: Verify the URL is correct and accessible
• Alternative: Try copying the article content manually

2. NotebookLM authentication failed

• Error: Auth/cookie error
• Solution: Run notebooklm login again
• Check: notebooklm status to verify

3. File upload failed

• Error: Invalid file or upload error
• Solution: Check if Markdown file was created correctly
• Verify: File path and permissions

4. Notebook creation failed

• Error: Rate limiting or API error
• Solution: Wait a few minutes and retry
• Alternative: Add to existing notebook with --notebook flag

Advanced Features

Add to Existing Notebook

If user wants to add to an existing notebook:

# Get list of notebooks
notebooklm list --json

# Use existing notebook ID
notebooklm source add /tmp/article.md --notebook <existing_notebook_id> --json

Batch Processing

For multiple WeChat articles:

# Create notebook once
notebooklm create "WeChat Articles Collection" --json

# Add multiple articles
for url in "${urls[@]}"; do
  # Fetch, save, upload to same notebook
done

Follow-up Actions

After successful upload, suggest:

For analysis:

notebooklm use <notebook_id>
notebooklm ask "What are the key insights?"
notebooklm ask "Summarize in 3 bullet points"

For content generation:

notebooklm generate audio "Create an engaging podcast"
notebooklm generate video "Make an explanatory video"
notebooklm generate quiz "Test understanding"

For research:

notebooklm source add-research "related topics" --mode deep
notebooklm ask "Compare with other sources"

Limitations

• WeChat articles only: Optimized for mp.weixin.qq.com URLs
• Text content: Focuses on text, images are preserved as links
• Public articles: Requires publicly accessible articles
• Rate limits: NotebookLM has rate limits on uploads

Troubleshooting

Problem: Article content is incomplete

• Cause: WeChat page uses JavaScript rendering
• Solution: web reader tool handles most cases, but some dynamic content may be missed

Problem: Chinese characters display incorrectly

• Cause: File encoding issues
• Solution: Ensure UTF-8 encoding when saving files

Problem: NotebookLM says "Processing" for too long

• Cause: Large articles take time to index
• Solution: Wait 1-2 minutes, then check status with notebooklm source list

Best Practices

1. Use descriptive notebook titles: Extract from article title or topic
2. Keep articles organized: Use separate notebooks for different topics
3. Verify uploads: Check notebooklm source list after upload
4. Clean up temp files: Remove /tmp files after successful upload
5. Handle rate limits: If uploads fail, wait 5-10 minutes before retry

Quick Reference

Example Usage

User: "Sync this WeChat article to NotebookLM: https://mp.weixin.qq.com/s/xxxxx"

Agent workflow:

1. Fetch article using web reader
2. Save to /tmp/article_title.md
3. Create notebook "Article Title"
4. Upload markdown file
5. Report success with IDs and next steps

Time estimate: 30-60 seconds total