Feature: AI Legal Background & Eviction Scraper
Overview:
Take first name, last name, and province from the ID.
Core Functionality:
Input: First Name, Last Name, Province Every check do the following:
summary: 1–2 sentence context (via OpenAI) sourceLink: direct link to case or listing JSON Output Example If Fail:
{
"name": "John Doe",
"province": "Ontario",
"match": "YES",
"summary": "Jane Doe was involved in a 2023 eviction dispute related to non-payment of rent.",
"sourceLink": "https://openroom.ca/dashboard/document/366832d6-1fae-420a-b051-c4f64f8f65eb/
JSON Output Example If Passed:
{
"name": "Adyan Tanver",
"province": "Ontario",
"match": "No",
"summary": "User Passed, no records found.",
"sourceLink": ""
Edge Case Handling:
If no results: return "match": "NO" with empty summary and null link If loading or site unreachable: return "match": "PENDING" or "ERROR" with status message Handles empty Airtable rows, invalid characters, and inconsistent data formatting
Tools:
Make.com – for automating trigger + data flow Crawl4AI – for scraping speed and reliability ChatGPT API – for natural language case summarization Airtable – as source and destination for input/output
Testing Interface:
Temporary frontend built in Next.js to manually trigger and test workflows before integrating with production at .
Logins:
Make.com
Airtable
Dump
Then for every search use crawler / ai to check these websites
Quebec → CanLII, SOQUIJ, TAL Alberta → Alberta Courts / QP Alberta
Use browserbase to get past, do later