Contact A-LUX

We reply within 15 minutes

Вы человек? 3 − 1 =

Or contact directly

Private LLM · Your data stays in-house

RAG System & AI Knowledge Base that answers from your own documents — with sources

Prices and terms valid as of June 2026

A-LUX builds retrieval-augmented generation (RAG) systems that read your contracts, regulations, 1C records and support history, then answer staff questions in seconds — every answer backed by a link to the exact source paragraph. No hallucinations, no data leaks.

WhatsApp
  • Answers cite the exact document and page — verifiable, not invented
  • Runs on a private LLM: your data never leaves your infrastructure
  • Connects to 1C, Kaspi, CRM, WhatsApp Business and file storage

A RAG (retrieval-augmented generation) system is an AI assistant that searches your company's own documents and answers questions with links to the exact source. A-LUX in Almaty builds these on a private LLM so your data stays in-house, with integrations to 1C, Kaspi and WhatsApp Business, starting from ₸1,200,000.

FAQ

Answers to common questions before ordering from A-LUX.

What is a RAG system in simple terms?

RAG stands for retrieval-augmented generation. Instead of relying on what an AI model "remembers", the system first retrieves the most relevant passages from your own documents, then asks the language model to write an answer grounded in those passages — and shows you the source. This makes answers accurate, verifiable and specific to your company.

How is this different from just using ChatGPT?

Public ChatGPT doesn't know your contracts, your 1C data or your internal policies, and it can invent plausible-sounding but wrong answers. A RAG knowledge base answers only from your documents, cites the exact source, and runs on a private LLM so confidential data never leaves your infrastructure.

Will our documents stay confidential?

Yes. We deploy an open-weight model on your servers or a dedicated VPS inside Kazakhstan. Documents, embeddings and queries are processed locally with encrypted storage and full access logging. Nothing is sent to a third-party API unless you explicitly approve it for non-sensitive tasks.

Which systems can the knowledge base connect to?

Commonly 1C, Kaspi order and payment data, amoCRM and Bitrix24, WhatsApp Business, Google Drive, SharePoint, Confluence and shared file storage. If your data lives in a database, a file or an API, we can index it.

How much does a RAG system cost?

Projects start from ₸1,200,000. The final price depends on the number of data sources, integrations (1C, Kaspi, WhatsApp), the volume of documents and whether you need on-premise hosting. After a short discovery call we give a fixed quote.

How long does it take to launch?

A working pilot covering your first knowledge domain is typically ready in about 4 weeks. Full rollout across all departments and integrations usually runs 2–3 months depending on data complexity.

Does it work in Kazakh and Russian?

Yes. The assistant handles questions and documents in Kazakh, Russian and English, and can answer in the language the user writes in — useful for mixed teams and customer-facing WhatsApp support.

How do you prevent the AI from making things up?

The model is constrained to answer only from retrieved content, every answer carries a source link, and we build an evaluation set of your hardest questions to measure citation accuracy. If the knowledge base has no answer, the assistant says so rather than guessing.

WhatsApp

Get a RAG system quote

Fill in the form — our manager will reply within 15 minutes. Free consultation and quote.

By submitting you agree to the processing of personal data