Katalove understands customer requests, finds the best-fit products, shows proof straight from the catalog, and collects quote requests. Even if today you sell through agents, phone, or email.
When the catalog contains hundreds of codes, variants and technical specs, the customer often doesn't know what to look for. They browse pages, send a generic email or give up. Katalove turns that catalog into a guided conversation.
Four steps. Tap a step to see what it produces.
Demonstrative example. The products shown are not a real catalog.
At 80 m³ with a 6-hour turnover you need roughly 13 m³/h. Here are two matched units, plus a larger one if you plan to extend the system.
We built Katalove Anchored Extraction, a proprietary method that turns even complex catalogues into structured, verifiable data.
Product codes, prices, variants, units of measure and categories are extracted according to rules defined together with your company, then checked and saved in the database. The AI seller can only use information anchored to the real data in your catalogue, which sharply reduces the risk of invented answers.
The catalogue becomes a reliable database. The AI becomes a competent salesperson.
Catalogs with hundreds of SKUs, variants and technical specs. You don't need to already have an e-commerce or a structured product feed.
Prices exclude VAT. The annual subscription equals ten monthly payments and is billed once a year.
At this stage plans are activated by talking to us: we look at the catalog and confirm the setup before starting.
A catalog holds your products, not your company. During setup we load the prices, write the selling conditions, the company profile and the questions the seller should ask, and check the products that were read wrong. We also put the seller on an address of yours, catalog.yourcompany.com. This work exists even when the PDF is perfect: from the second catalog it costs less, because company and conditions are already written.
The final setup price is confirmed after analyzing the material.
The limits shown refer to catalogs with an ordinary structure. Scanned PDFs, complex tables, numerous variants, multiple price lists or image-product associations may require a specific assessment.
No. A document chatbot searches for words in a file and returns a piece of text. Katalove rebuilds the catalog as a set of products, codes and attributes, asks questions to understand what's needed, compares alternatives and collects a quote request.
PDF, CSV and Excel. Alternatively you can have your website read, if the products are public.
No. If the pages are images with no selectable text, there's nothing to read. We tell you right after the analysis, counting how many pages are silent, and usually all you need to do is ask the supplier for the print PDF.
Item codes are read from the PDF text along with their position on the page, not asked from a model: that way none are lost. Images are extracted and linked to products, and we have you confirm any uncertain matches.
Every important answer shows the catalog page it comes from, or the page on your website. You open it and read it.
It only answers based on what it has read in your sources, and when a piece of data isn't there it says so instead of guessing. It never infers prices or availability.
Yes. A catalog can be protected with a PIN and stays accessible only to those you choose.
Yes, by uploading the new version. The sales agent starts working from the updated material.
A conversation is a session started by the same visitor on the same channel. The session ends after 30 consecutive minutes of inactivity, and messages exchanged within the same session don't count as separate conversations. A very long session gets closed and a new one begins.
We notify you before any interruption and you can move up to the higher plan.
Not yet available. Today the sales agent can be reached from the widget on the website, from the Katalove page and from QR codes. Any Meta costs, number and templates wouldn't be included anyway.
Because preparing a technical catalog is real work: extraction, verification of uncertain elements, image checking. The price is confirmed after seeing the material.
Not yet. Integrations are evaluated case by case on Enterprise projects.
On a compatible catalog, a first demo is generated in a few minutes. Final publication depends on verifying the extracted data and the sales conditions you want to give the sales agent.
Each company's data is separate and reserved catalogs can't be reached without a PIN. For details on processing and retention, the privacy policy is linked at the bottom of the page.
Yes, anytime. Setup already done isn't refundable.
Send us your catalog or your site address. We'll tell you what we can read and what we can't, for free.
No card required. We check for free whether your catalog is compatible.
On compatible catalogs a first demo can be generated in a few minutes.
Do you have a website with products? Import it →