Reference Implementation: n8n Workflows¶
This directory contains reference workflow exports for common automations.
1. Email Intake to Paperless¶
File: email-to-paperless.json (Mocked)
Triggers: IMAP node.
Actions: Filter by attachment, POST to Paperless API.
2. Paperless Tag to Google Calendar¶
File: tag-to-calendar.json (Mocked)
Triggers: Webhook from Paperless.
Actions: Get OCR text, Send to LLM node, Create GCal event.
How to use¶
- Import the JSON into your n8n instance.
- Configure credentials for IMAP, Paperless, and Google.
- Update the LLM model ID to your local Ollama or OpenAI instance.
Local LLM Node Configuration¶
To use local models within n8n, use the OpenAI Chat Model node with a custom base URL.
1. Connection via Ollama¶
- Resource: Chat Model
- Model:
qwen3-coder-next:latest(or your preferred model) - Base URL:
http://<your-server-ip>:11434/v1 - Authentication: None (unless configured with a proxy)
2. Connection via LiteLLM¶
LiteLLM can act as a unified gateway for multiple local models.
- Base URL: http://<litellm-ip>:4000
- Model: The model name as defined in your config.yaml
Token-Efficiency Tip¶
Disable Memory (History) in the LLM node for one-off extraction tasks to keep the context window small and save local compute resources.