Home Finder

What it is

Home Finder is a Python-based real-estate search and aggregation tool for Romanian property listings. It combines several source-specific scrapers with configurable filters, duplicate removal, and a small local browser interface so a search can be reviewed without manually opening every source.

The problem and business value

Property searches are fragmented across multiple listing sites, each with different markup and result shapes. Home Finder brings those sources into one reviewable result set, applies the same location, property-type, price, and surface filters, and keeps the original listing link for the next step. The value is a faster, more consistent first pass over a property search.

The repository does not claim listing completeness, seller verification, investment advice, or any commercial result.

What it does today

• Registers scrapers for nine Romanian real-estate websites: imobiliare.ro, storia.ro, olx.ro, homezz.ro, imoradar24.ro, romimo.ro, galaxyimob.ro, kingimobiliare.ro, and imopedia.ro.
• Applies configurable filters for location, property type, price, rooms, useful area, land area, and construction stage.
• Normalizes result columns, removes tracking parameters from links, and drops duplicate listings.
• Runs source workers in isolated processes with timeouts and progress events so one unavailable site does not block the complete search.
• Provides a local HTTP UI and command-line runner, with CSV and Excel exports in the local output directory.

Integrations and data flow

The command-line runner and local UI pass a filter set to the source registry. Each scraper returns a tabular frame, the parent process normalizes and filters the records, canonicalizes listing links, and returns a deduplicated result set. The browser UI receives job status and partial source results through local JSON endpoints.

The project uses Selenium and Chrome for browser-based acquisition. Source pages and their terms can change independently; the code therefore treats selectors, availability, and access permission as ongoing maintenance work rather than a permanent guarantee.

AI or automation workflow

There is no AI workflow in the current repository. Automation is deterministic browser scraping, filtering, normalization, and export. The local UI starts a bounded background job; it does not make property decisions or send listings to third parties.

Lead, email, notification, or collaboration flow

No lead capture, email delivery, or external notification flow is shipped. Results are written for local review and can be exported as CSV or Excel for a separately authorized next step.

Security and privacy boundaries

The tool is local-first and does not include a hosted data store or provider credentials. Listing pages may contain personal or sensitive information, so output files should be kept private and source terms and applicable law should be checked before any reuse. Rate limiting, anti-bot behavior, and site-specific access rules remain operational concerns for every source.

What is coming next

The next work is selector maintenance, deterministic tests for each source, clearer per-source error reporting, and stronger rate-limit and legal-policy documentation. A durable shared service or public listing distribution path would require a separate privacy, terms, storage, and operational review.

Current activity or evidence from the project

The repository currently contains the scraper registry, filter and deduplication pipeline, local HTTP UI, worker isolation, and export paths. The README and source modules support these claims; there is no evidence here of production hosting, complete source coverage, listing freshness, or business outcomes.