Built a local real-estate lead qualification demo in n8n

Hi everyone — I built a self-initiated portfolio demo for qualifying buy and rent enquiries in n8n. This is a learning project, not client work.

The flow accepts an enquiry through a form or webhook, validates and normalizes it, optionally uses a local Ollama model to extract fields from free text, parses the model output with a fallback, calculates an explainable readiness score, checks synthetic listing data, and stores the lead in an n8n Data Table. It can return a mock notification preview or send a formatted Gmail alert.

I verified structured buyer and renter examples, free-text extraction with local Ollama, a fallback path, lead upsert behavior, and a synthetic Gmail notification. Telegram is present but disabled. The workflow runs on my local n8n instance, so there is no public interactive demo URL.

The score measures how complete and ready the enquiry is based on stated details; it is not a housing eligibility or conversion decision. Listing data and test enquiries are synthetic, and a human agent reviews the next action.

Portfolio case study and verification notes: Real Estate Lead Qualification

The portfolio page includes the full walkthrough and these two PNG proofs. The workflow screenshot shows structure; the email screenshot shows synthetic test content and does not by itself prove delivery metadata.

n8n workflow canvas

Synthetic Gmail test email

Building it helped me practice webhooks, validation, JSON parsing, branching, local LLM connections, storage, and email formatting. I’d welcome feedback: what would you improve in the qualification or listing-match steps?

1 Like