Multilingual News Curator AI Agent

What This Agent Does

This intelligent workflow automatically monitors RSS feeds for relevant content, processes articles through AI-powered analysis, and organizes the results in Google Sheets while keeping your team informed via Slack. The agent runs on a schedule, filtering articles by keywords, rewriting content for clarity, extracting location data, translating between multiple languages, and formatting everything into a structured database—all without manual intervention.

Key benefits include:

• Save 10+ hours weekly on manual content monitoring and data entry
• Never miss important articles with automated keyword filtering and real-time notifications
• Break language barriers with automatic translation to English and Chinese
• Centralize intelligence in Google Sheets for easy team access and analysis

Perfect use cases:

• Media monitoring and competitive intelligence gathering
• International news tracking for global teams
• Content curation for research or marketing purposes
• Automated industry trend analysis

Who Is It For

This workflow is ideal for:

• Marketing teams who need to track industry news and competitor content across multiple languages
• Research analysts gathering intelligence from international sources
• Content curators building databases of relevant articles for newsletters or reports
• PR professionals monitoring media coverage and industry developments
• Business intelligence teams tracking market trends and competitor activities

Whether you're a solo entrepreneur or part of a larger organization, this agent eliminates the tedious work of manually checking feeds, translating content, and organizing information.

Required Integrations

RSS Feed Reader

Why it's needed: This integration fetches articles from your target RSS feeds, serving as the data source for the entire workflow.

Setup steps:

1. Identify the RSS feed URL(s) you want to monitor
2. Verify the feed is accessible by opening it in a browser
3. Note the feed structure (most standard RSS feeds work automatically)

Configuration in TaskAGI:

• No API key required for public RSS feeds
• The RSS node accepts any valid feed URL
• Private or authenticated feeds may require additional headers

Slack

Why it's needed: Slack provides real-time notifications when articles are skipped (due to keyword filtering) or successfully processed, keeping your team informed without checking spreadsheets.

Setup steps:

1. Go to api.slack.com/apps and click Create New App
2. Choose From scratch and name your app (e.g., "TaskAGI RSS Monitor")
3. Select the workspace where you want notifications
4. Navigate to OAuth & Permissions in the sidebar
5. Under Bot Token Scopes, add these permissions:
   • chat:write (to send messages)
   • chat:write.public (to post in public channels)
6. Click Install to Workspace and authorize the app
7. Copy the Bot User OAuth Token (starts with xoxb-)

Configuration in TaskAGI:

1. Go to Integrations in your TaskAGI dashboard
2. Select Slack and click Add Integration
3. Paste your Bot User OAuth Token
4. Test the connection by sending a test message
5. Invite the bot to your desired channel by typing /invite @YourBotName in Slack

Anthropic (Claude AI)

Why it's needed: Claude powers three critical AI operations: rewriting articles for clarity, extracting location information, and translating content between languages. This integration is the intelligence engine of your workflow.

Setup steps:

1. Visit console.anthropic.com and create an account
2. Navigate to API Keys in your account settings
3. Click Create Key and give it a descriptive name (e.g., "TaskAGI RSS Workflow")
4. Copy the API key immediately (it won't be shown again)
5. Note your usage limits and pricing tier

Configuration in TaskAGI:

1. Go to Integrations → Anthropic
2. Click Add Integration and paste your API key
3. Test the connection with a simple prompt
4. The workflow uses claude-sonnet-4-5-20250929 model for optimal balance of speed and quality

Cost considerations: Each article processes through Claude three times (rewrite, extract, translate). Monitor your Anthropic usage dashboard to track costs.

Google Sheets

Why it's needed: Google Sheets serves as your centralized database, storing all processed articles with structured data including original content, rewrites, locations, and translations.

Setup steps:

1. Create a new Google Sheet or use an existing one
2. Set up column headers in the first row (e.g., Title, Original Content, Rewritten, Location, English Translation, Chinese Translation, Date, Source URL)
3. Copy the full sheet URL from your browser
4. Go to console.cloud.google.com
5. Create a new project or select an existing one
6. Enable the Google Sheets API for your project
7. Navigate to Credentials → Create Credentials → Service Account
8. Name your service account (e.g., "TaskAGI RSS Writer")
9. Grant it Editor role
10. Click on the created service account and go to Keys tab
11. Click Add Key → Create New Key → JSON
12. Download the JSON key file

Configuration in TaskAGI:

1. Go to Integrations → Google Sheets
2. Upload the JSON key file or paste its contents
3. Share your Google Sheet with the service account email (found in the JSON file as client_email)
4. Grant Editor permissions to the service account
5. Test the connection by attempting a write operation

Configuration Steps

1. Schedule Trigger Setup

Configure when your workflow runs automatically:

• Interval: Set to 3600 seconds (1 hour) for hourly checks, or adjust based on your needs
• Daily: Use 86400 for once-daily monitoring
• Custom: Calculate seconds for your preferred frequency (e.g., every 4 hours = 14400)

Recommendation: Start with hourly checks and adjust based on feed update frequency.

2. Fetch RSS Feed Configuration

Point the workflow to your content source:

• Feed URL: Enter your RSS feed URL (e.g., https://example.com/feed.xml)
• Item Limit: Set to 10 to process the 10 most recent articles per run
• Output: The node provides items array containing article data

Data available from RSS:

• title - Article headline
• description - Article summary or full content
• link - Original article URL
• pubDate - Publication timestamp

3. Filter Keywords Configuration

Determine which articles to process:

• Condition Type: Select Contains or Does Not Contain
• Field to Check: Use [[nodes.4418.items.0.description]] to check article content
• Keywords: Enter comma-separated keywords (e.g., technology, AI, automation)
• Case Sensitivity: Typically set to false for broader matching

True path: Articles matching keywords proceed to processing False path: Non-matching articles trigger a skip notification

4. Notify Skipped Configuration

Keep track of filtered-out content:

• Channel: Enter your Slack channel (e.g., #rss-monitoring)
• Message Template:

Article skipped (no keyword match): [[nodes.4418.items.0.title]]
URL: [[nodes.4418.items.0.link]]

5. Rewrite Article Configuration

Transform articles for clarity and consistency:

• Model: claude-sonnet-4-5-20250929 (pre-configured)
• Prompt: [[nodes.4418.items.0.description]] with instructions to rewrite for clarity
• Max Tokens: Set to 2000 for comprehensive rewrites
• Temperature: Use 0.7 for balanced creativity and consistency

Output: Stores rewritten content in [[nodes.4421.output]]

6. Extract Location Configuration

Identify geographical references:

• Model: claude-sonnet-4-5-20250929
• Prompt Template:

From this article, extract all location mentions (cities, countries, regions):
[[nodes.4418.items.0.description]]

Return only location names, comma-separated.

• Max Tokens: 500 (locations are typically brief)

Output: Comma-separated location list in [[nodes.4422.output]]

7. Translate Content Configuration

Make content multilingual:

• Model: claude-sonnet-4-5-20250929
• Prompt:

Translate this Japanese article to English and Chinese. 
Provide both translations clearly labeled.

Article: [[nodes.4418.items.0.description]]

• Max Tokens: 3000 (translations need more space)
• Temperature: 0.3 (lower for more literal translations)

Output: Contains both English and Chinese translations

8. Format Data Configuration

Structure data for Google Sheets:

• Function Type: Select JavaScript or Python
• Input Variables: Map all previous node outputs
• Code Template:

return {
  title: nodes[4418].items[0].title,
  original: nodes[4418].items[0].description,
  rewritten: nodes[4421].output,
  location: nodes[4422].output,
  translations: nodes[4423].output,
  url: nodes[4418].items[0].link,
  date: new Date().toISOString()
};

9. Save to Google Sheets Configuration

Write processed data to your spreadsheet:

• Sheet URL: Paste your Google Sheet URL (e.g., https://docs.google.com/spreadsheets/d/YOUR_SHEET_ID)
• Row Data: Map formatted data fields to columns
• Append Mode: Ensure true to add new rows without overwriting

Column mapping example:

• Column A: [[nodes.4424.title]]
• Column B: [[nodes.4424.original]]
• Column C: [[nodes.4424.rewritten]]
• Column D: [[nodes.4424.location]]
• Column E: [[nodes.4424.translations]]
• Column F: [[nodes.4424.url]]
• Column G: [[nodes.4424.date]]

10. Notify Slack Configuration

Celebrate successful processing:

• Channel: Same as skip notification channel
• Message Template:

✅ New article processed!
Title: [[nodes.4424.title]]
Locations: [[nodes.4424.location]]
View in sheet: [YOUR_SHEET_URL]

Testing Your Agent

Initial Test Run

1. Disable the schedule trigger temporarily to prevent automatic runs during testing
2. Click Run Workflow in the TaskAGI interface
3. Select Manual Trigger to execute immediately
4. Watch the execution flow in real-time

Verification Checklist

After RSS Fetch:

• ✓ Verify articles are retrieved (check node output)
• ✓ Confirm article descriptions contain content
• ✓ Note the number of items fetched

After Keyword Filter:

• ✓ Check which path was taken (true/false)
• ✓ Verify keyword matching logic works correctly
• ✓ Confirm skip notification appears in Slack if false path

After AI Processing:

• ✓ Review rewritten content for quality and clarity
• ✓ Verify locations are accurately extracted
• ✓ Check translations for both English and Chinese
• ✓ Ensure translations maintain original meaning

After Google Sheets Save:

• ✓ Open your Google Sheet and verify new row appears
• ✓ Check all columns are populated correctly
• ✓ Confirm formatting is readable
• ✓ Verify URLs are clickable

After Slack Notification:

• ✓ Confirm success message appears in designated channel
• ✓ Check message formatting and readability
• ✓ Verify all data references display correctly

Expected Results

A successful test run should:

• Complete in 30-90 seconds depending on article length
• Create one new row in Google Sheets per processed article
• Send one Slack notification (either skip or success)
• Show green checkmarks on all executed nodes
• Display no error messages in the execution log

Troubleshooting

RSS Feed Issues

Problem: "Failed to fetch RSS feed" error

Solutions:

• Verify the feed URL is accessible in a browser
• Check for CORS restrictions (some feeds block automated access)
• Try adding a User-Agent header: Mozilla/5.0
• Ensure the feed uses standard RSS/Atom format

Problem: No items returned from feed

Solutions:

• Increase the item limit in the RSS node
• Check if the feed has been recently updated
• Verify the feed isn't empty or deprecated

Anthropic/Claude Issues

Problem: "Rate limit exceeded" error

Solutions:

• Reduce workflow frequency (increase schedule interval)
• Implement a delay between AI calls using a Wait node
• Upgrade your Anthropic plan for higher rate limits
• Process fewer articles per run

Problem: Poor translation or extraction quality

Solutions:

• Adjust the temperature setting (lower = more consistent)
• Refine your prompts with more specific instructions
• Increase max_tokens if responses are cut off
• Add examples in your prompts for better results

Google Sheets Issues

Problem: "Permission denied" error

Solutions:

• Verify the service account email has Editor access to the sheet
• Check that the sheet URL is correct and complete
• Ensure the Google Sheets API is enabled in your Google Cloud project
• Confirm the JSON key file is valid and not expired

Problem: Data appears in wrong columns

Solutions:

• Review column mapping in the Save to Google Sheets node
• Ensure column order matches your sheet structure
• Check that data references use correct node IDs
• Verify the Format Data node outputs match expected structure

Slack Issues

Problem: "Channel not found" error

Solutions:

• Invite the bot to the channel: /invite @YourBotName
• Use channel ID instead of name (found in channel details)
• Verify the bot has chat:write permissions
• Check that the channel is public or bot is a member

Problem: Messages not formatting correctly

Solutions:

• Use Slack's mrkdwn formatting syntax
• Escape special characters in dynamic content
• Test message templates with static text first
• Review Slack's message formatting documentation

Workflow Logic Issues

Problem: Keyword filter not working as expected

Solutions:

• Test with simple, single keywords first
• Check case sensitivity settings
• Verify the field reference points to article content
• Use OR logic by creating multiple condition branches

Problem: Workflow stops mid-execution

Solutions:

• Check execution logs for specific error messages
• Verify all required fields are populated in each node
• Ensure data flows correctly between nodes
• Test each node individually to isolate the issue

Next Steps

After Successful Setup

Immediate actions:

1. Enable the schedule trigger to start automatic monitoring
2. Monitor the first 24 hours to ensure stable operation
3. Review processed articles in Google Sheets for quality
4. Adjust keyword filters based on initial results

Optimization Suggestions

Improve accuracy:

• Refine your keyword list based on false positives/negatives
• Enhance AI prompts with specific instructions for your industry
• Add additional filtering conditions for better targeting

Scale your monitoring:

• Add multiple RSS feed nodes for different sources
• Create separate workflows for different content categories
• Implement priority routing based on article importance

Enhance notifications:

• Add rich formatting to Slack messages with article summaries
• Create different notification channels for different content types
• Implement digest notifications (daily summary instead of per-article)

Expand data collection:

• Add sentiment analysis using additional Claude prompts
• Extract key entities (people, organizations, products)
• Generate article summaries or key takeaways
• Track trending topics over time

Advanced Usage Tips

Multi-language support:

• Add more translation targets by duplicating the translate node
• Create language-specific Google Sheets for different audiences
• Implement automatic language detection for mixed-feed sources

Data enrichment:

• Integrate with web scraping to fetch full article content
• Add image extraction and storage
• Connect to CRM or marketing automation platforms
• Implement automatic categorization or tagging

Analytics and reporting:

• Create Google Sheets formulas for trend analysis
• Build dashboards using Google Data Studio
• Export data to BI tools for deeper insights