Troubleshooting¶
Common issues and solutions for the DeepTagger n8n node.
Installation Issues¶
Node doesn't appear after installation¶
Symptoms: Can't find DeepTagger in node list
Solutions:
-
Restart n8n completely
-
Verify installation
Should show:n8n-nodes-deeptagger@x.x.x
-
Check n8n logs
-
Clear browser cache
- Hard refresh:
Ctrl+Shift+R
(Windows/Linux) orCmd+Shift+R
(Mac)
npm install fails¶
Error: npm ERR! code EACCES
Solution:
# Fix npm permissions
mkdir ~/.npm-global
npm config set prefix '~/.npm-global'
export PATH=~/.npm-global/bin:$PATH
npm install n8n-nodes-deeptagger
Error: npm ERR! network
Solution:
Authentication Issues¶
Error: "No API key found in request"¶
Cause: Credential not configured or not selected
Solutions:
- Create DeepTagger API credential:
- Go to Credentials → New Credential
- Search "DeepTagger API"
-
Enter API Key and Base URL
-
Select credential in node:
- Click on DeepTagger node
-
Choose credential from dropdown
-
Verify API key is not empty in credential settings
Error: "Unauthorized" or "401"¶
Cause: Invalid or expired API key
Solutions:
- Verify API key in DeepTagger dashboard
- Login to deeptagger.com
- Go to Settings → API Keys
-
Check key is active
-
Regenerate API key:
- Create new key in dashboard
- Update credential in n8n
-
Test again
-
Check for extra spaces:
- API key should have no leading/trailing spaces
- Copy directly from dashboard
Error: "Invalid base URL" or "Network Error"¶
Cause: Incorrect base URL or connectivity issue
Solutions:
-
Verify base URL is exactly:
(No trailing slash!) -
Check for typos:
http
vshttps
-
/api/v1
vs/api/v2
-
Test connectivity:
-
Check firewall (self-hosted only):
- Ensure n8n can reach deeptagger.com
- Whitelist in corporate firewall if needed
Execution Issues¶
Error: "Project not found" or "404"¶
Cause: Incorrect Project ID
Solutions:
- Verify Project ID format:
- Must start with
fo_
-
Example:
fo_1759714105892
-
Get correct Project ID:
- Go to deeptagger.com/das/fos
- Open your project
-
Copy ID from URL bar
-
Check project exists:
- Log in to DeepTagger dashboard
- Verify project is not deleted
No data extracted (empty result {}
)¶
Cause: Untrained project or wrong document type
Solutions:
- Verify project is trained:
- Go to DeepTagger dashboard
- Open project
-
Check it has 3+ training examples
-
Test in dashboard first:
- Upload same document in DeepTagger UI
- Verify it extracts correctly
-
If it doesn't work in UI, train more examples
-
Check document quality:
- Is the PDF readable?
- Is text selectable (not scanned image)?
-
Try a different document
-
Verify document type matches:
- Invoice project should process invoices
- Don't send receipts to invoice project
Error: "Binary data not found"¶
Cause: Previous node doesn't output binary data
Solutions:
- Check previous node outputs binary:
- HTTP Request (binary mode)
- Webhook (file upload)
- Google Drive / Dropbox
-
Email (with download attachments)
-
Verify Binary Property name:
- Check previous node's output
- Common values:
data
,attachment0
,file
-
Set DeepTagger Binary Property to match
-
Use Move Binary Data node:
- Add between nodes
- Move binary from custom property to
data
Text input doesn't work¶
Cause: Wrong Input Type or missing text
Solutions:
-
Set Input Type to "Text"
-
Provide text content:
- Can be hardcoded for testing
-
Or dynamic:
{{$json["body"]}}
-
Check text is not empty:
- Add Code node before to debug
- Log:
console.log($json)
Performance Issues¶
Execution timeout¶
Cause: Large files or slow network
Solutions:
- Increase timeout in n8n settings:
- Settings → General → Execution Timeout
- Default: 120 seconds
-
Try: 300 seconds (5 min)
-
Optimize file size:
- Compress PDFs before processing
-
Split large documents
-
Check network:
- Test API speed outside n8n
- Contact support if consistently slow
Rate limit errors¶
Cause: Too many requests
Solutions:
- Add delays in batch workflows:
- Wait node between iterations
-
1-2 second delay recommended
-
Reduce concurrency:
- Process files sequentially, not in parallel
-
Use Loop node instead of multiple branches
-
Check your plan limits:
- Login to DeepTagger dashboard
- View API usage and limits
- Upgrade plan if needed
Data Quality Issues¶
Extracted data is inaccurate¶
Cause: Needs more training or document variation
Solutions:
- Add more training examples:
- Minimum: 3 examples
- Recommended: 5-10 examples
-
Include variety of document layouts
-
Review annotations:
- Check training examples in dashboard
- Ensure selections are accurate
-
Re-annotate if needed
-
Check document quality:
- Low-resolution images reduce accuracy
- Scanned documents may need OCR cleanup
Missing fields in output¶
Cause: Field not present or not trained
Solutions:
- Check if field exists in document:
- Some invoices may not have all fields
-
Handle missing data in workflow
-
Verify field is in project schema:
- Dashboard → Project → Schema
-
Add missing fields if needed
-
Train with examples containing that field:
- Annotate at least 3 examples with the field
Wrong data type (string instead of number)¶
Cause: DeepTagger returns all data as strings
Solutions:
-
Convert in n8n expression:
-
Use Code node for complex transformations
Workflow Issues¶
Duplicate entries in database¶
Cause: Workflow running multiple times
Solutions:
- Implement duplicate check:
- Query database before insert
-
Check unique field (invoice number)
-
Use idempotency keys:
- Add unique ID to each execution
-
Skip if ID already processed
-
Archive processed items:
- Move processed emails to folder
- Mark files as processed
Workflow fails on some documents¶
Cause: Inconsistent document format or extraction failures
Solutions:
- Enable "Continue on Fail":
- Node settings → Continue on Fail
-
Workflow continues even if node fails
-
Add error handling:
- IF node after DeepTagger
- Check:
{{$json["error"] === undefined}}
-
Success branch vs Error branch
-
Log errors:
- Send failed items to error database
- Review and retrain with failed examples
Docker/Kubernetes Issues¶
Node not available in Docker¶
Cause: Package not installed in container
Solutions:
-
Install at runtime:
-
Use custom Dockerfile:
-
Persist installation:
- Mount
/home/node/.n8n
volume - Installation persists across restarts
Kubernetes pod restarts lose node¶
Cause: No persistent volume
Solutions:
-
Use PersistentVolumeClaim:
-
Or use init container (see Installation)
Getting Additional Help¶
If these solutions don't resolve your issue:
1. Check Documentation¶
2. Search GitHub Issues¶
github.com/deeptaggerhq/n8n-nodes-deeptagger/issues
- Search existing issues
- Check if your problem is already reported
- See solutions from community
3. Open a New Issue¶
When creating an issue, include:
- n8n version:
n8n --version
or from Settings - Node version:
node --version
- DeepTagger node version: Check in package.json or npm list
- Installation method: npm, Docker, Kubernetes
- Error message: Full error with stack trace
- Steps to reproduce: Workflow JSON if possible (remove sensitive data!)
- What you've tried: List solutions you've attempted
4. Contact Support¶
For n8n node issues: - GitHub: github.com/deeptaggerhq/n8n-nodes-deeptagger/issues
For DeepTagger API/platform issues: - Email: support@deeptagger.com - Dashboard: Help widget (bottom-right)
For n8n platform issues: - Forum: community.n8n.io - Docs: docs.n8n.io
5. Community Resources¶
- n8n Community Forum
- DeepTagger Discord (coming soon)
- Stack Overflow (tag:
n8n
+deeptagger
)
Debug Mode¶
Enable debug logging in n8n for detailed error information:
# Self-hosted
export N8N_LOG_LEVEL=debug
n8n start
# Docker
docker run -e N8N_LOG_LEVEL=debug n8nio/n8n
Check logs for detailed API requests/responses.
Reporting Bugs¶
Found a bug? Help us fix it!
- Verify it's a bug (not expected behavior)
- Check it's reproducible (happens every time)
- Create minimal test case (simplest workflow that shows the bug)
- Open issue on GitHub with:
- Title: Brief description
- Environment: n8n version, OS, installation method
- Steps to reproduce
- Expected behavior
- Actual behavior
- Workflow JSON (if applicable)
- Screenshots (if helpful)
Thank you for helping improve DeepTagger!