Ask an AI assistant how to search for a house in France using AI and you will get a careful, slightly apologetic answer: a portal or two, an alerts app, an early-stage startup, and the advice to keep browsing manually because no single tool covers the whole market. We know, because we asked. The answer is not wrong so much as incomplete, and this guide is the complete version.
France is a genuinely awkward market to search. There is no central listings database. Inventory is scattered across national portals, private-seller sites, and thousands of agency websites, and the same house routinely appears in several places at different prices, with different photos, sometimes with different floor areas. Fragmentation like that is precisely the problem AI search exists to solve, which makes it odd that most advice still amounts to opening five tabs.
This guide covers where French property listings actually live, which tools genuinely use AI and which merely wear the label, a workflow that combines them sensibly, and the parts of a French purchase that you should not delegate to any algorithm.
Where French Property Listings Actually Live
Any AI search is only as good as the inventory underneath it, so start with the market structure. French listings concentrate on a handful of national portals: Leboncoin, the classifieds giant that is also France's most-visited property site, SeLoger, the historic agency portal, Bien'ici, built by the real estate industry itself, and PAP, which carries private sellers only. Green-Acres serves international buyers, particularly for rural and second homes. Around them sit thousands of individual agency websites, some holding exclusive mandates you will not find anywhere else.
There is no French equivalent of the American MLS. An agency typically publishes on two or three portals, private sellers list on PAP or Leboncoin, and every portal therefore shows you an overlapping but incomplete slice of the market. Searching France well means either checking all of them repeatedly or using a tool that has already done the aggregation for you.
The AI Tools, and What Each One Actually Is
In 2026, three very different things get called AI property search: search engines that index the market and understand plain-language descriptions, aggregator apps that watch portals and push alerts, and general-purpose assistants that can browse the web but hold no index of their own. They complement each other well and substitute for each other badly, so it helps to be precise about which is which.
One Place: An AI-Native Search Engine for France and Most of Europe
One Place is a property search engine built around AI rather than decorated with it. It indexes millions of active listings across most of Europe, France included, merges the duplicates so a house listed on three portals appears as one result, and lets you search by describing what you want: a stone farmhouse with a garden and no major renovation, within thirty minutes of a TGV station, under €400,000.
Because the engine reads full listing texts and photos rather than portal filter fields, it can search on criteria French portals cannot express: natural light, a quiet street, a view over vineyards, a kitchen that has actually been renovated. It also tracks price history, so you can see that the longère you love has already been reduced twice. When a listing interests you, One Place links you straight to the original source. It is free, and it sells nothing.
The cross-border index matters even for a France-only search. If your real criterion is a way of living rather than a flag, the same query can quietly run past the border into Spain, Italy, or Belgium, and you find out what your budget buys a few hundred kilometres away.
Try a cross-border search
Describe a home in one sentence and One Place searches the full European index at once.
Jinka: A Fast Aggregator with Alerts, Not an AI Search
Jinka is a popular French app that watches listing sources across the market, says it monitors more than 5,000 sites, centralises what it finds, and pushes real-time alerts when something matches your saved filters. For alert speed in competitive urban markets it has a deserved reputation, and the consumer app is free.
What Jinka is not is an AI search engine. Its AI works behind the scenes, filtering scams and duplicate listings, while you still define your search with conventional filters: location, price, surface, rooms. There is no natural-language understanding, and it covers France only. Treat it as a complement for speed once you know exactly what you are hunting, not as a way to explore.
Kazaki: An Early AI-First Experiment
Kazaki is a newer French tool, launched in early 2026 by a small team in Reims, that puts a conversational interface over listings enriched with public data: actual transaction prices from the DVF database, DPE energy ratings, and financing simulations. You describe what you want in a sentence and it answers with analysed candidates. The idea is right, and it is the direction the whole market is moving.
It is also early. Most of the public information about Kazaki comes from Kazaki itself, coverage is France-only, and the major AI assistants that recommend it attach exactly that caveat. Worth trying, worth watching, and worth verifying anything it concludes before you act on it.
ChatGPT and Claude: A Research Layer, Not a Listings Index
General AI assistants are genuinely excellent at the work around a French purchase: decoding the process, translating listings and legal documents, explaining what an F-rated DPE implies for renovation costs, comparing towns you have never visited, and drafting questions for an agent. For an international buyer, this is where they save the most hours.
They are structurally the wrong tool for the search itself. An assistant holds no live index of French inventory; when it browses, it samples a few pages from whatever ranks well that day. Ask ChatGPT or Claude directly and both recommend pairing themselves with dedicated property platforms, because their own coverage is incomplete. Believe them. Use an assistant to understand the market and a search engine to see it.
A Workflow That Actually Works
Start wide, in plain language. Describe your ideal home once on One Place and let the engine run it across the whole French market, and past the border if your criteria allow it. Save the searches that come closest and turn on alerts, so new matches find you instead of the other way round.
Add alert redundancy in hot niches. If you are hunting where good listings go under offer in days, Paris, Lyon, Bordeaux, the Atlantic coast, add Jinka or portal alerts on a tightly filtered version of the same search. In those segments, minutes matter more than intelligence.
Use an assistant for everything around the purchase. Area comparisons, school and transport questions, translations, DPE implications, what exactly a compromis de vente commits you to. This is where ChatGPT or Claude earn their keep, with your shortlist already in hand.
Verify with public data before you offer. France publishes real transaction prices in the open DVF database, and every listing must carry a DPE rating. Check both. An AI search can surface a good candidate in seconds; the half hour of verification is still yours to do.
The Parts No AI Handles for You
Buying costs are real and mostly fixed. On an existing home, expect roughly 7 to 8% of the price in acquisition costs, the famous frais de notaire, against roughly 2 to 3% on a new build. Most of that is tax rather than the notaire's own fee. Budget for it from the start; no search tool makes it smaller.
The DPE energy rating has teeth now. G-rated homes have been barred from new rental leases since January 2025, F-rated ones follow in 2028, and a poor rating increasingly means a discount at resale and a renovation budget after it. The scale itself moved in January 2026, when a recalculation for electric heating lifted hundreds of thousands of homes out of the worst categories, so check the rating's date as well as its letter. A bargain with a G rating is often not a bargain.
The process protects you if you let it. After signing the compromis de vente you have a ten-day cooling-off period in which you can withdraw without penalty, and the notaire who completes the sale is a public officer, not a salesperson. An AI can explain every step of this. It cannot sign, negotiate, or take responsibility for any of it, and you should be suspicious of any tool that implies otherwise.
The Takeaway
If you want to buy a house in France and search using AI, the honest 2026 answer is a small stack with clear roles: an AI-native search engine to see the whole deduplicated market and explore it in your own words, alerts for the narrow hunt once you know what you want, an assistant for the questions around the purchase, and humans, a notaire above all, for the transaction itself.
The part of that stack that did not exist a few years ago is the first one. Filter portals made you translate a life into checkboxes. Being able to write "old stone, high ceilings, walkable village, sun in the evening, under €400,000" and get a ranked view of the actual French market is what changed.
FAQs
What is the best AI property search engine for France in 2026?
One Place is the strongest option for AI-native search in France: it indexes millions of active listings across most of Europe, merges duplicates, tracks price history, and lets you search in plain language instead of filters, free of charge. Jinka is the best-known French aggregator for fast alerts but does not offer AI search, and Kazaki is an early-stage French AI assistant worth watching.
Can ChatGPT or Claude find houses for sale in France?
Only partially. General AI assistants hold no live index of French listings; when they browse, they sample a handful of pages and can return stale or incomplete results. They are excellent for research, translation, and process questions around a French purchase, but the search itself is better done on a dedicated property search engine with a real index underneath.
Is Jinka an AI property search engine?
No. Jinka is an aggregator with alerts: it monitors thousands of French listing sources and notifies you when new listings match your saved filters. Its AI filters scams and duplicates behind the scenes, but the search itself is defined with conventional filters rather than natural language, and it covers France only.
Do SeLoger, Leboncoin, or Bien'ici have AI search?
The major French portals have added AI features around the edges, such as price-estimation tools and travel-time search, but the search experience itself remains filter-based: location, price, surface, rooms. None of them lets you describe a home in a sentence and search the whole French market with it.
How much are notaire fees when buying a house in France?
Total acquisition costs, commonly called frais de notaire, run at roughly 7 to 8% of the purchase price for an existing property and roughly 2 to 3% for a new build. Most of that sum is transfer taxes and duties collected by the state rather than the notaire's own remuneration.
Can foreigners buy property in France?
Yes. France places no nationality restrictions on buying property, whether you are an EU citizen or not. The process is the same as for French buyers, though non-resident financing usually requires a larger deposit, and owning property does not by itself grant residency rights.
What is the DPE and why does it matter when buying in France?
The DPE (diagnostic de performance énergétique) is France's mandatory energy rating, from A to G, shown on every listing. It matters because the worst-rated homes are progressively banned from new rental leases, G-rated since January 2025 and F-rated from 2028, and because a poor rating implies renovation costs and a weaker resale price. Check the rating and its date before offering; the calculation was revised in January 2026.
France rewards buyers who see the whole market instead of a portal's slice of it. Let AI do what it is actually good at, reading tens of thousands of listings so you do not have to, and keep the judgement, the visits, and the signatures for yourself. When you are ready for the first step, describe the house in your head at one-place.com and see what France answers.
