Kloufi gathers Algerian property listings; DZ-Immobilier turns them into services — estimate, check, finance, visit. Here is, without jargon, what artificial intelligence can change in the market, what already exists, what is in preparation, and what it will never do in place of a professional.
Figures as of October 24, 2026. Source: Kloufi, kloufi.com/statistiques; last figure: DZ-Immobilier’s counter on Kloufi’s index, updated every night. Listings — not properties, not sales.
A listing is only a starting point. Between it and an informed decision there is cleaning, comparing, explaining — and, at the end, always a person.
Listings scattered across many websites, in several languages, with very different ways of writing a price.
The starting pointA listings search engine: it brings together listings from Algerian websites, sorts them into categories, ties them to a wilaya, converts prices, flags prices worth checking — and always links back to the original listing.
Search and dataFrom this market data and the tax authority’s (DGI) reference grid, DZ-Immobilier computes price ranges and always gives a confidence level — or no price at all when data is missing.
AnalysisNotary, expert, bank, agent: what binds you stays in their hands. Estimate, check a deed, finance, visit: our services prepare the decision, they do not take it.
DecisionKloufi and DZ-Immobilier are two separate platforms. DZ-Immobilier cites Kloufi as the source of its market data, by agreement between the two platforms, which share the same technical management.
Eleven possible uses, from the most mature to the most exploratory. For each one: how it works, the technique actually used, who benefits, where we stand — honestly, including when the answer is “not yet”.
In general, a property is compared with comparable ones (area, type, size, condition) and a reference grid, and a range is deduced: the more data, the more recent and the closer, the more the range can narrow. Machine learning can refine this calculation when there is enough data.
Individuals who sell or buy, agencies, developers, banks.
DZ-Immobilier’s online estimate is a rules engine: a reference price per m², adjusted to the property. That price combines asking prices from listings (Kloufi data), the tax authority’s DGI grid and values produced by language models, now frozen. The result: a range and a confidence level; when data is missing (“expert required”), no price is shown.
An asking price in a listing is not a sale price. The confidence level is not an accuracy rate, which we have not measured. An estimate is neither a certified valuation nor a guaranteed price.
Tolerant search matches what you type with what listings contain: synonyms, typos, language variants. A semantic search would instead compare the meaning of a sentence with the meaning of listings.
Buyers and tenants, agencies.
On Kloufi, search understands French, Arabic, Darija, Kabyle and Tifinagh, tolerates typos and knows former names of municipalities, thanks to a vocabulary maintained by the team. By voice, the sentence is transcribed by the browser; place, price and property type are interpreted on the device and you confirm. Search by meaning remains an idea.
An engine that “understands” can be wrong: it must show the criteria it used and let you correct them.
A language model writes a description from the characteristics entered, then a translation; the professional reads and approves.
Agencies, developers, individuals posting a listing.
An idea: we do not offer it today. It will only be opened with mandatory human review.
A generated text can invent a detail: no description should be published without a person’s approval.
Computer vision can measure sharpness, detect duplicates and recognise the type of room to help check listings.
Agencies, buyers, listing-review teams.
An idea. Today only simple checks exist (resolution, file size, exact duplicates); image recognition proper is not in service.
A score is only an aid: the decision to publish remains human.
Panoramic shots are assembled into a navigable tour, hosted by DZ-Immobilier, with no app to install.
Buyers (including the diaspora), developers, agencies.
Available: properties published on DZ-Immobilier offer a 360° tour. AI could one day help improve images or produce a floor plan; that is not the case today.
A virtual tour replaces neither an on-site visit nor a check of the documents.
A conversational assistant could answer using the listing’s information, gather the visitor’s need and prepare the request for the professional.
Agencies, developers, individuals.
An idea: not offered today. To be opened only with controlled sources and a clear human handover.
An assistant must never invent a price or an availability.
Rules read the title and description (advance payment, contact imposed outside the platform…) and compare the price with that of comparable properties in the municipality.
Buyers, agencies, banks.
Kloufi shows a “Price to check” or “Warning” badge on listings whose price or text clearly departs from the market.
It is a signal, not an accusation or a proof; no detection is complete.
Automatic text reading can extract the fields of a deed and flag inconsistencies; a person then checks them against official registers.
Individuals, notaries and lawyers (as a reading aid), banks.
DZ-Immobilier offers a deed check: one request, then a report in French and Arabic reviewed by the team, presented as “an alert service, not an authority”.
It is neither a certificate of authenticity, nor an expert opinion, nor a lawyer’s or notary’s advice: the deed is the notary’s matter, advice is the lawyer’s.
Cleaned listings are aggregated and indicators are computed by area and property type, stating the sample size and the date.
Agencies, developers, banks, institutions.
Kloufi publishes regularly recalculated statistics (listings online, wilayas, property types, transactions). DZ-Immobilier uses them to place a property and offers a simulation of the tax authority’s DGI reference grid.
Asking prices are not sale prices. We are preparing fuller, sourced and dated indicators before publishing them.
Comparing a programme with nearby supply, tracking enquiries by programme and by type, decision support.
Developers and their sales partners.
In preparation: developer registration is open; the tools to follow programmes and enquiries are being designed.
Off-plan sales follow a strict legal framework: the tool informs, it does not replace the notary.
Transparent calculators apply the rates, terms and incomes entered; an AI could then help explain the results.
Individuals, banks.
Available: loan simulator, monthly payment and borrowing capacity calculation. Rates and debt rules are those shown on the site, to be confirmed with your bank.
A simulation is not a loan offer.
“Available” does not mean “AI”: many of our services rely on rules and transparent calculations. We state the technique used so that everyone can judge.
An automatic estimate is not a magic number: it is a range, with a confidence level. Here is the general principle: move the slider to see the effect of the number of comparable properties. With too little data, it is better to put no figure at all.
Educational illustration of the general principle: fictitious values and thresholds. It is neither the formula nor the thresholds of our engine, which starts from a reference price adjusted by rules, and our estimate’s confidence level is not an accuracy rate.
Too few comparables: no price is given, an expert is required.
Writing listings, searching by image, estimating a price: foreign portals already use it, and some even publish their errors. Four reference points, each with its date and its source.
ImmoScout24 and Homegate (Switzerland) extend AI generation of a listing’s title and description to private individuals. The company announces a saving of about 14 minutes per property for professionals (company figure).
SourceRightmove presents five AI features launched within five months: search by keywords read from photos and texts, virtual staging, renovation cost estimate, an assistant in search and in ChatGPT.
SourceZillow announces a national median error of 6.9% across more than 104 million off-market homes. Fictitious example: at 7%, a property worth 30 million DA is off by more than 2 million DA at the median.
SourceZillow winds down its home-buying business Zillow Offers: a $304 million inventory write-down in the third quarter of 2021, and an executive cites the unpredictability of forecasting home prices.
SourceIn Algeria, actual sale prices are not published: an estimate therefore remains, by nature, more uncertain than elsewhere. That is why we give a range, a confidence level and the date of the data. These foreign measures do not transpose as such to Algeria; they give an order of magnitude of the uncertainty.
Services open to everyone, designed for the Algerian market and available in French, Arabic, English and Spanish.
A price range and a confidence level for your property, with a report.
Look up the DGI reference value per m² by period, wilaya, municipality and type of property.
Have a deed reviewed before you sign: a report flags what should alert you.
Simulate a monthly payment and your borrowing capacity.
Walk through properties from your phone, wherever you are.
Notaries, lawyers, experts, agencies, developers: sign up, the team validates your request.
Our project: give agencies, developers and legal and valuation professionals a space worthy of a modern portal, with AI where it brings real value. Steps come in order; no date is announced until it can be kept.
Agencies, developers, notaries, lawyers, experts and individuals go through the same circuit, with validation by the team.
One page per professional, with their listings, programmes, team — and a clear mention of what has been checked, with its date.
Every contact reaches the professional’s space, with follow-up and an announced response time.
What matters to an agency is not what matters to a developer: everyone sees their own priorities.
Residences, buildings, units, availability: a reliable view, consistent with what the public sees.
Comparables, indicators and reports, always dated and accompanied by their sample.
Suggest a listing text or a translation: the professional reads, corrects and approves before any publication.
In a market where trust is scarce, transparency is our advantage.
On this page, every figure carries its date and source, and every use its status. What is indicative is called indicative.
AI proposes and prepares. What binds you — a deed, a valuation, a loan, a contract — stays in the hands of the notary, the expert, the bank and you.
Each form says why data is requested and collects explicit consent, a requirement of Algerian law 18-07.
Better one clean, dated and sourced data point than ten uncertain ones: an asking price stays an asking price.
Journalists, real-estate players, institutions, partners: what you need to understand and relay our approach, with dated figures and ready-to-quote texts.
DZ-Immobilier is an Algerian real-estate services platform in four languages (French, Arabic, English, Spanish). It offers an indicative estimate as a range with a confidence level, a simulation of the DGI reference grid, financing simulators, documentary deed checks and support for sellers, buyers and the diaspora. Its market data comes from Kloufi.
Kloufi is an Algerian listings search engine, specialised by category (real estate, cars, multimedia, home appliances, jobs). It gathers listings from Algerian websites, locates them by wilaya, flags prices worth checking and always links back to the original listing. Its search understands French, Arabic, Darija, Kabyle and Tifinagh, and tolerates typos.
These figures describe published listings, not properties or completed sales. They change every day: always quote the date.
Use the logos as they are, without modification. Kloufi’s name, logo and texts belong to Kloufi.
Kloufi is an Algerian listings search engine: it gathers listings from Algerian websites and always links back to the original listing; it neither publishes nor sells listings. DZ-Immobilier cites Kloufi as the source of its market data, by agreement between the two platforms, which share the same technical management.
Not in the usual sense of the term. Today, Kloufi’s search relies on rules and dictionaries maintained by its team, and voice interpretation happens on the device. Kloufi itself says its machine learning is in an observation phase. We prefer to say it clearly.
A rules engine starts from a reference price per m² — which combines asking prices from listings (Kloufi data), the tax authority’s DGI grid and values produced by language models, now frozen — then adjusts it to the property: area, type, size, condition. The result is a range and a confidence level; when data is missing, no price is shown. It is neither a valuation nor a sale price.
Every figure on this page is dated and sourced, and we say what it counts: listings online, listings indexed — never properties or sales. Volumes change from one day to the next, because source websites publish and remove listings continuously.
Yes, in a limited way, and we prefer to say so. Reference price values were produced by language models, then frozen; articles on our blog are written with the help of a language model. Where AI would be useful — describing or translating a listing, an assistant — it will only be offered with human review, and its status will be shown on this page.
Each form states the purpose of collection and collects explicit consent, a requirement of Algerian law 18-07. For any question about your data, use the Contact page.
We move in steps, in order, and we announce a date only when it can be kept. Registration is open; the rest is presented above with its status.
Use the “Contact the press team” button: we reply within 48 hours (business days).
Try the online estimate and judge the method for yourself. And if you are a professional, tell us what is missing.
Page updated on October 24, 2026.