Baza is a digital platform for real estate agents that combines a mobile app, backend, API, search infrastructure, and a listing parsing microservice. The product was created for the Ukrainian real estate market, with the first launch in Dnipro and further scaling planned for other cities.
We developed the product from scratch: from UX/UI and mobile apps for iOS and Android to backend architecture, parser, DevOps infrastructure, and QA. The main focus was to give agents a single workspace for viewing the market, finding properties for clients, and maintaining their own database.
Challenge
Real estate agents work daily with many scattered sources: listing websites, personal notes, chats, colleagues’ contacts, and informal collections. Because of this, it is hard to see the market as a whole, and finding the right property for a client often turns into manually monitoring dozens of pages.
The key problem was not only collecting listings. We needed to create a mobile tool that helps agents quickly assess the entire available market: owner listings, offers from other agents, new publications, personal database, statuses, client collections, and exclusive properties.
Our role
We were responsible for the full product development cycle. We designed the UX/UI, implemented the mobile apps with React Native and TypeScript, developed the backend and API, built the parser in Python, configured geodata storage in PostGIS, search through Elasticsearch, Docker containerization, infrastructure, testing, and deployment processes.
The architecture was also created from scratch. This made it possible to split the product into clear technical areas from the start: mobile client, API, property database, search index, listing parser, push notifications, and infrastructure for continuous service operation.
Product solution
At the center of the product is the “Baza” section, where the user sees properties from different sources in a unified format. Listings can be filtered by source, author type, price, floor, building floors, area, and other parameters needed by agents during property selection.
Each property has a detailed card with photos, description, characteristics, price, and contact information. An agent can add personal notes, change the status, save a contact, move the property to their own database, or quickly share it with a client through messengers and social networks.
Agent’s personal database
The “My database” section works as a personal workspace. A user can add properties from the general database, create their own records, edit data, assign statuses, remove outdated items, search, filter, and group properties into folders.
Folders make it possible to maintain collections for specific clients, districts, budgets, deal types, or any other internal workflow logic. A separate Favorites section was implemented for properties that users need to return to quickly.
Filter channels and fast reaction
Filter channels were implemented for continuous market monitoring. An agent saves a set of criteria, and the system tracks new properties that match those parameters. When such a property appears, the user receives a push notification.
This scenario is critical for the real estate market: only a few minutes pass between a listing being published on a website and appearing in the app. An agent can quickly see a new offer and be the first to suggest it to a client.
Exclusive property exchange
A separate part of the product is the “Eksy” tab for exclusive properties. Agents can add their own offers and make them visible to other platform users. This turns the app not only into a search tool, but also into a professional channel for exchanging information between market participants.
Data, parsing and search
The most complex technical part was the parsing microservice. It continuously monitors different websites with different page structures, collects thousands of listings, converts them into a unified format, and passes them to the system for further processing.
The parser extracts address, price, area, floor, building floors, description, images, and other property characteristics from unstructured text. We also solved deduplication, address normalization, and extraction of parameters that are important for search and filtering.
PostGIS is used for geodata, while Elasticsearch powers fast search and filtering across a large volume of listings. This makes it possible to work with thousands of properties and quickly find relevant options by a combination of parameters.
Backend and infrastructure
The backend handles users, properties, personal databases, folders, statuses, filter channels, push notifications, exclusive property exchange, and data synchronization between the mobile apps. The API was designed as a clear contract between the mobile client and the server side.
The infrastructure was deployed on DigitalOcean using Docker. The services were designed with continuous parser operation, API stability, data persistence, and the ability to scale to new cities and a growing number of listings in mind.
Result
As a result, the startup received a digital product launched from scratch: mobile apps for iOS and Android, UX/UI, backend, API, listing parser, search infrastructure, push notifications, DevOps, and QA processes.
For real estate agents, Baza became a tool that helps them see the market more completely, quickly search for properties among listings from owners and other agents, maintain their own database, react to new publications within minutes, and exchange information with colleagues more conveniently.
