AI Lab

Real demos, with their limits. Keys and models stay on the server.

01 · rag_search

Document search (RAG)

Query this site's published content: the demo retrieves the most relevant passages and their sources.

How to use: Type a question in natural language.
Limits: Limited to the portfolio's published content. Without an embedding model, search uses BM25.

Result

Type a question in natural language.

02 · text_analysis

Summary, sentiment, key points and entities of a text.

How to use: Paste a text (article, email, customer review).
Limits: Do not enter personal or confidential data. Indicative results, to be checked.

Result

Paste a text (article, email, customer review).

03 · extraction

Turns free text into structured data (JSON).

How to use: Paste an email or request containing contact or order details.
Limits: Missing fields are returned empty: the model must not invent anything, but always check the output.

Result

Paste an email or request containing contact or order details.

04 · classification

Routes a message to the right team automatically.

How to use: Paste a customer message.
Limits: Fixed categories set in the admin. Confidence estimated by the model.

Result

Paste a customer message.