Each result is a link to the relevant page.
Vocabulary No results Word, term, topic…Patents — Alias Integrator
Three patent applications have been filed with the French National Institute of Industrial Property (INPI). They lock the priority date of the mechanisms at the core of Alias — and therefore of your investment.
Technical details remain confidential until the applications are published. Extensions are being prepared: a United States application and a European patent application.
Application FR2612512 — filed September 15, 2026
The problem. Business repositories (lexicons, catalogs, nomenclatures) are classically coded for a single domain: adding a category or switching domains requires modifying the code and migrating the database. Every adaptation costs a development.
What the application enables. A single data structure serves arbitrarily different domains — tourism, documentation, real estate, etc. — without code modification or schema migration: everything is driven by declarative configuration. Switching domains means replacing a configuration and data, never rewriting the system.
The same database serves any business domain without rewriting: the declarative configuration drives the structure, the consistency and the maintenance of the repository.
Application FR2612514 — filed September 15, 2026
The problem. A conventional conversational AI generates the most probable answer — it can distort a price, a date, a name. Known approaches measure the error after the fact or inject documents into the context, but no displayed value is guaranteed to match the source.
What the application enables. The factual fragments of an answer — a price, an availability, a reference — are determined by directly querying your data, independently of any probabilistic generation. Where a conventional AI can make things up, ours is accurate by construction. And all processing runs locally, with no third-party service.
The displayed factual values come directly from the operator's database, read-only — they cannot diverge from the source.
Application FR2612515 — filed September 15, 2026
The problem. Querying a database requires developer skills. Existing visual builders expose the query language; generic no-code tools assemble pre-designed blocks without checking that the query matches the intent. As a result, training an AI on your own data remains reserved for technicians — or outsourced to a third party.
What the application enables. A non-programmer composes a correct query by drag and drop, without ever seeing the query language: the interface continuously checks that the composition is valid, rephrases it in plain language for confirmation, and tests beforehand that the answer actually exists in the database. Training the AI becomes a business gesture, not a development.
The operator composes, checks and validates their query entirely in plain language; the system guarantees it returns what was asked.
Framework
These three applications are provisional applications filed with INPI in the name of the creator and publisher of the ecosystem — owner of the code and of all intellectual property (see Terms). The provisional application locks the priority date while leaving 36 months for completion and international extensions: the framework provided by French intellectual property law to secure an innovation during its structuring phase. Extensions to the United States and the European patent are being prepared.
A crowdfunding campaign is being prepared to support the completion and the international extensions. To be informed of its opening — or to talk beforehand — write to the creator / publisher: show the email address.