Turn messy search into matched filters

Type a messy renter query. The Python NLP layer extracts intent, ranks dummy listings, and explains why each property matched.

Results per page

Match strictness

Loading dataset...

Parsed intent

{}

Verified source data should stay separate from NLP guesses and search-only fields.

sourceportal, listing URL, agent id, last seen timestamp
pricepcm, weekly rent, deposit, bills included
propertyaddress, postcode, lat/lng, type, beds, baths, size, floor, build age
availabilityavailable date, tenancy length, furnishing, pets, student friendly
amenitiesgarden, balcony, parking, gym, concierge, lift, step-free access
proximitynearby stations, landmarks, commute estimates
nlpclean text, derived tags, negative tags, confidence
searchtext blob, embedding id, indexed timestamp

hard filtersrent ceiling, room type, location, station/landmark proximity, exclusions
soft prefsamenities, floor level, build age, fuzzy lifestyle terms
vector soft fitTF-IDF vector score boosts lifestyle wording after hard filters stay in control
explainabilityevery match stores hard checks, soft scores, exclusions, and reasons