What We Do

In most companies 1C works in isolation. Data on goods, stock, counterparties and documents lives inside it, while everything outside — the website, requests from messengers, reports for management — is filled in by hand. Employees move the same numbers from one system to another, waste time and make mistakes. Every discrepancy has to be tracked down manually.

We solve this through the OData API — the native 1C:Enterprise protocol that lets you read and write data directly, in real time, with no intermediate Excel files and no customizations that break configuration updates.

Moving goods between warehouses with two-sided confirmation

The client's challenge: goods constantly move between warehouses and divisions. Previously this meant paper waybills, calls of "you shipped it — we didn't receive it," stock discrepancies, and back-dated postings in 1C, when it's already unclear who accepted what.

The same approach works for any business where data from 1C is needed beyond it:

If you have 1C and at least one task your staff do "by hand because there's no API" — chances are it can be automated through OData.

Modern language models (Claude and other LLMs) can do what once required a team of people: understand text, extract data from unstructured documents, answer questions, compose responses. But "adding AI" doesn't mean dropping a chat box on your website. The value appears when the model works with your data and is built into your process.

We deploy AI not as a toy but as a working tool: with access to the right data, with oversight, and with a clear payoff.

Face recognition service for medical organizations

The client's challenge: identify a patient from a photo at check-in — fast, with no manual database search and with data separated between medical organizations.

What we did: an asynchronous API built on FastAPI: face detection in the image, building a FaceNet embedding and nearest-vector search via FAISS. Records are linked to patients and organizations, access is protected with JWT tokens, and each clinic's data is isolated from the others.

We build AI solutions on the Claude API and choose the model to fit the task — taking data-confidentiality requirements into account. For sensitive data we use modes with limited retention.

Off-the-shelf boxed products rarely fit a company's real processes. You end up either bending your processes to someone else's logic or paying licenses for features no one uses. And the data still lives across several disconnected systems, between which employees move information by hand.

We build systems for specific tasks on Django — a reliable, proven framework that handles both small internal portals and platforms for thousands of users. The key difference in our solutions: they aren't isolated but connected to your 1C and other systems, so data isn't duplicated.

The Adami HR platform for a manpower company

The client's challenge: an outsourcing company with a large headcount. Data on each worker already exists in 1C — hiring, transfers, details. But managing personnel calls for a separate HR system, and rolling it out raises the question: re-enter every employee by hand? For a staff of thousands that's weeks of work and inevitable errors.

What we did: Adami is an HR platform that doesn't store employee data separately but pulls it directly from 1C. When a worker is hired and registered in 1C, they automatically appear in Adami with all their details. There's no need to enter the same person twice — 1C remains the single source of truth for personnel, while Adami adds everything 1C lacks on top: process management, roles, approvals, personnel analytics.

The principle is always the same: we don't create yet another system you have to duplicate data into. We build a web layer on top of what you already have.

Every company has operations that repeat dozens of times a day by the same rules: take a request, check stock, send a notification, enter data, reconcile a register. These tasks require no qualification, but they eat up working time and slow the process down — the customer waits while a manager frees up.

We automate such operations in two ways: scripted automation (RPA) where systems need to be linked together, and chatbots where the interaction is with customers and employees.

We work on official APIs (the WhatsApp Business Cloud API, Telegram) — which means stability and no risk of being blocked, unlike "grey" emulation-based solutions.

In Kazakhstan, the EDS is the foundation of legally binding document workflow. But in many systems a "signature" amounts to a pretty badge image in a PDF with no real cryptography behind it. Such a document can't be verified and has no legal force. And where an EDS is genuinely needed, the process often stays manual: export, sign in a separate program, upload back.

We embed a real cryptographic signature right into your web systems — at the level of the CMS and PAdES standards, via NCALayer and NCANode.

This is a competency that few on the market have. If your processes involve documents that need signing, we'll move them to a fully electronic form — with real legal force, not an imitation.

Wherever there is a cash desk and a flow of visitors, several disconnected systems usually coexist: one sells tickets, another takes payments, a third issues receipts, a fourth controls the turnstile. Data between them is reconciled by hand, and any failure at the seams means either lost revenue or a double charge to the customer.

We build ticketing and payment systems as a single whole: from online payment to the turnstile opening — one loop with a shared database, fiscalization and reporting. Payment operations are processed idempotently: pressing “pay” twice or a connection failure never creates a second charge.

Ticketing and payment systems for the Almaty and Shymkent zoos

The client's challenge: the city zoos needed the whole cycle automated: ticket sales and refunds, cashless payments, entry control and transparent reporting — instead of queues at the ticket office and manual revenue reconciliation.

What we did: end-to-end systems for two sites: ticket sales and refunds by category, QR code generation and validation at the turnstiles, payments via Kaspi and Halyk, fiscalization of every operation via Webkassa, parking tickets and camera events. For management — reporting by sales, cashiers and payment methods.

Money is the one area where “almost works” doesn't count. That's why our payment solutions are built on idempotent operation processing, protection against duplicate charges and reconciliation with fiscal receipts.