Tabnine Not Working? Fix Common Issues

Completions not appearing in VS Code

1. 1
   Check the status bar icon: Look for the Tabnine icon at the bottom of VS Code. A yellow warning means Hub is not connected; grey = loading; green = working.
2. 2
   Check the Output log: View → Output → select "Tabnine" from dropdown. Error messages here tell you exactly what is wrong.
3. 3
   Conflict with GitHub Copilot: Both tools use VS Code's inline suggestion slot. If Copilot is enabled, its suggestions may suppress Tabnine. Try disabling Copilot temporarily with "GitHub Copilot: Disable Completions" from the Command Palette.
4. 4
   Workspace trust restriction: In a Restricted workspace, extensions including Tabnine are blocked. Check the shield icon in the status bar and set trust to allow Tabnine.
5. 5
   Reset the engine: Run "Tabnine: Reset Tabnine Engine" from the Command Palette. This restarts the local language server without reinstalling.

Tabnine Hub not connecting

Tabnine Hub runs locally on port 5555 and manages cloud model access. If it fails to connect:

• Port conflict: check if something else is using port 5555. On Windows: netstat -ano | findstr 5555
• Firewall blocking: Tabnine Hub needs outbound HTTPS to api.tabnine.com and update.tabnine.com. Allow them through corporate proxy/firewall.
• Antivirus quarantining Hub: Some antivirus software blocks Tabnine's local server binary. Add Tabnine's install directory to the exclusion list.
• Reinstall: Uninstall the extension, delete ~/.tabnine (Linux/Mac) or %APPDATA%\Tabnine (Windows), and reinstall.

AI Chat not available

The free tier includes local model completions only. For AI Chat, you need Tabnine Pro ($12/month or $9/month annually). Upgrade at app.tabnine.com/profile.

High CPU / slow IDE

Tabnine can be CPU-intensive on older machines, especially with large codebases:

• Switch to a smaller model: In Tabnine settings, select "Small" model — much lower resource usage with similar accuracy for most code.
• Limit to specific languages: In the extension settings, disable Tabnine for file types where you don't need completions (JSON, Markdown, YAML).
• Disable on large files: Set a file size limit in settings — Tabnine will skip files above that threshold.
• Use local model only: Set Tabnine to "Basic" mode (offline, local model). Lower quality than cloud models but zero CPU overhead from Hub.

JetBrains IDE (IntelliJ, PyCharm, WebStorm)

• Plugin not loading: Go to Settings → Plugins → verify Tabnine is enabled and not in "Disabled" or "Error" state. Restart the IDE.
• Completions not appearing: Check Settings → Editor → General → Code Completion → ensure "Show suggestions as you type" is enabled.
• Slow start: JetBrains with Tabnine can be slow on first load — the plugin indexes your project on startup. Give it 60–90 seconds before reporting slowness.
• Plugin incompatibility: After a major IDE update, the Tabnine plugin may need an update too. Check for plugin updates in Settings → Plugins → Updates.