A city leisure platform with an AI assistant
A parser collects the city’s venues and event listings, and an AI assistant plans an evening from a request in plain words. Table bookings, tickets, chat and business dashboards live in one system.
Discuss a similar project
- 103
- screens across 10 user flows
- 4
- roles: user, organizer, venue, admin
- 5
- data sources in the parser: 2GIS, Sxodim, Ticketon, Threads and Instagram
Challenge
City listings are scattered across 2GIS, ticketing services and event sites, so people pick their evening by hand. The client needed a catalog of venues and events that keeps itself up to date, evening picks tailored to each person, and a sales channel for organizers and venues: bookings, tickets and reviews.
What we built
A catalog parser with adapters for 2GIS, Sxodim, Ticketon, Threads and Instagram, running on a schedule. The same event from different sources is merged into one card, addresses become points on the map, and an LLM fills in missing categories and descriptions. The parser publishes nothing on its own: items wait in a moderation queue, and moderators get an alert when a source breaks.
An AI assistant: people describe what they want in their own words, and the assistant replies and builds an evening route with times. The model sees only the published catalog and can refer only to its cards, so it never invents places. Its picks take into account the date and time in the city, distance, and the interests, atmosphere and going-out time from the user’s profile. The same LLM writes the “Why people come here” text for each card strictly from the facts, and the model itself can be switched with a single setting.
The user journey: a personal home screen and a “For you” feed that explains each pick, a map, venue and event pages with reviews, table bookings, and tickets with ticket tiers, promo codes and QR codes. Go Together finds company for an event based on shared interests. Once both people agree, a chat opens with real-time messages and read receipts. A morning digest sends picks for the day, and push notifications remind people about bookings and events.
Business and admin dashboards. Organizers manage events, ticket sales, promo codes and mailings, while venues take bookings, maintain their menus and reply to reviews. Everything new goes through moderation. Admins manage roles, complaints and blocks, parser sources and the LLM prompt, and every action is logged.
Product screens
Result
The client received the app, business dashboards, parser and AI assistant as one system. The parser keeps the city catalog up to date, and moderators only review and publish.