EN TR
Get a quote
View all services
Web software

One app carries the brochure pages and the request queue.

System architecture

The pieces and their boundaries are settled before code.

Enterprise & commercial software

Several institutions, one system, records kept apart.

Mobile app

One codebase, both stores: iOS and Android.

Software consulting

We take a running system over and improve it.

AI consulting

AI only where the problem actually calls for it.

Get a quote
EN TR

Enterprise & Commercial Software · Esenyurt

Enterprise & Commercial Software for Esenyurt businesses

On the tutoring floors around Saadetdere and Yesilkent in Esenyurt, years of printed question booklets sit on a shelf. The answer key is a separate sheet. When a branch wants to run its own mock exam, booklet and key are matched by hand, and a lost key makes a booklet unusable. The question cropping tool we wrote for Vipkurum removes that split — a page is divided into single questions, and the answer from the key is written into each question's own record.

Get a quote Tell us about your project

Enterprise & Commercial Software — Esenyurt

Private education is dense in Esenyurt and runs across branches, so one brand can be teaching in several buildings at once. The work closest to that texture is Vipkurum's question cropping tool. Where it stops is clear: accounting, purchasing and staff records are not in its scope, and a multi-branch retailer's till flow is not something we can show. What we can describe is one conversion — how a question printed on paper becomes a record, and where its answer is stored. Python and OpenCV handle the cropping; Gemini and OpenAI models handle the matching and the writing of new questions. Cropped images and question records live in MongoDB, and the image service connects through Node.js to content management on the Laravel side.

  • A booklet page is split into single questions by OpenCV, so the boundaries come from image processing and that decision is never left to a language model's reading.
  • The answer from the key is bound to its question and written into the cropped question's record, which ends the answer key's life as a separate document.
  • Existing questions serve as examples for writing new ones in the same form, and every new question is produced together with its own answer key.

Why Algoriel

  • Many institutions, one system

    Vipkurum runs seven institutions on the same system while each one keeps its records to itself.

  • Adding one adds a boundary

    Each Vipkurum institution signs in at its own domain under its own brand, so adding one does not mean a second installation.

  • Each branch runs its own work

    A branch marks its own exam sheets, so one broken device does not stall every institution at once.

  • Each role, its own view

    Timetable, exams, students, parents, payments and reporting sit on one platform, and who can see what is settled up front.

What we build

Enterprise & Commercial Software

  • Each institution signs in through its own domain

    At Vipkurum, seven education institutions open under their own address and their own brand, and the name of the software never appears on a user's screen.

  • No reinstallation for a new institution

    Adding one more institution means writing a definition. No second copy of the system is stood up, and the maintenance stays in one place.

  • Office and field on the same record

    AlgoKlima's office panel and field app read and write the same record. A technician closes the job where there is no signal, and the entry reaches the office panel once the connection returns.

  • Operational and commercial records close together

    At AlgoCar the service record and the commercial record sit in the same system, so the cost of a job and its revenue are not reconciled across two programs.

  • One institution's data stays closed to the next

    Student, parent and payment records stay inside the boundary of the institution they belong to. That boundary is how the system is set up, not a view switched off on a screen.

  • One technical footing we can maintain for years

    Nothing goes into your system that the team maintaining it does not already know. AlgoKlima and AlgoCar are our own products, built on that same footing, and we work on it every day.

Work like this

Enterprise & Commercial Software · Case study

Vipkurum question cropping — the booklet becomes a pool, the answer joins the record

The tool we wrote for Vipkurum separates two jobs. First cropping: OpenCV finds the boundaries of the questions on a booklet page, and no language model interprets anything there. Then matching: Gemini and OpenAI models bind the answer from the key to its question and write it into the cropped question's own record. The key stops being a separate document and the booklet turns into a pool read one question at a time. A third door opens: existing questions become examples for writing new ones, each produced with its own key. Images and records sit in MongoDB; the Python service connects through Node.js to Laravel content management.

Read the case study →

Working with Esenyurt teams, remotely

We have no office, so work for this city runs remotely and we do not offer on-site support. What was done, and why it was done that way, stays in writing. The project pages mark which builds came out of a partnership.

  • The first release ships as a working piece; extras follow later.
  • Every client we know of is in Turkey.
  • Decisions stay in writing; progress shows in working screens.

Frequently asked

  • Our booklets are printed. Where does the work start?

    At the image of the page. In the cropping step OpenCV finds the boundaries of each question, so nothing has to be retyped or cut by hand. Every question that comes out becomes its own record, and the later steps are written onto that record.

  • Does the software invent the answer key?

    No. The source is your own key, and the model only binds each answer to the question it belongs to. Writing new questions is a separate step, and there the question and its key are produced together and kept in the same record.

  • What happens when a crop comes out wrong?

    The two jobs were split on purpose. Image processing does the cropping, the language model does the interpretation. A boundary error stays an image processing problem and never leaks into the model's answer, so which step needs fixing is never in doubt.

  • Could our branches share one question pool?

    The pool is what this tool produces: cropped questions and their records accumulate in MongoDB and connect to content management through Laravel. How institutions and branches are separated is a scoping decision of its own, and we do not hand it over as a ready template.

Enterprise & Commercial Software in other cities

The other cities where we publish this service are below, grouped by province.

Other services for this location

How many institutions and branches should we plan your system around?

Get a quote or open the service