HomeServicesPricingAboutCase studiesContactFAQBlog
Back to case studies

MAIstr

An AI knowledge manager your know-how cannot walk out of

View live
MAIstr - the interface for training an AI agent by uploading video, audio, a document or text

About the project

MAIstr came out of a real problem in our family business. The knowledge of experienced people lives in their heads, in notes and in email, and when someone leaves, how things are done leaves with them. MAIstr lets know-how be captured four ways: upload a document, record it by voice, film a video, or simply type it. The AI processes the input, extracts what matters and stores not just the information itself but the relationships between it and what is already in the system. That is what makes the answers accurate and relevant rather than merely similar to the question. An employee then asks in plain language and gets an answer immediately, at any hour. Data security is treated as a priority, because what goes into the system is the most valuable thing a company has. The project also includes the maistr.cz website, built on Next.js and Tailwind. It is not there just to describe the product but to demonstrate it: the MAIstr assistant runs on the page, connected to a knowledge base about the product, so visitors try the feature instead of reading about it. Behaviour on the site is measured with Google Analytics and Microsoft Clarity, and wherever people hesitate we adjust both the site and the onboarding into the app. The project is developed under my own name and is in a pilot phase, taking on its first customers.

Project information

Client
Own product
Platform
Web application
Status
In development
Development time
Ongoing, pilot phase
Year
2026

Technologies used

Check out the stack I used for this project

Frontend
ReactViteNext.jsTailwind CSS
Backend
PythonGeminiClaude
Database
QdrantPostgreSQL
Tools
Google AnalyticsMicrosoft Clarity

Key features

Knowledge captured four ways: document, voice, video, text

Inputs processed by AI into a knowledge base automatically

Relationships between pieces of information preserved, not isolated snippets

Questions in plain language, answers immediately and at any hour

Specialised agents for individual areas of the business

Overview of unanswered questions as a prompt to fill knowledge gaps

Vector search across the knowledge base

Next.js marketing site with the MAIstr assistant running live on it

Visitor behaviour measured with Google Analytics and Microsoft Clarity

Challenges

  1. 01

    Knowledge is scattered across the company and often exists only in people's heads

  2. 02

    Capturing know-how without costing an experienced person hours of writing

  3. 03

    Similarity search returns fragments without context

  4. 04

    The most valuable company data goes in, so security cannot be an afterthought

  5. 05

    Explaining the value of an AI product to a website visitor in seconds

  6. 06

    Combining several models and data stores into one working whole

Solutions

  1. 01

    Four input routes including voice and video: describing a procedure out loud is enough

  2. 02

    AI processes, sorts and prepares the input into the knowledge base on its own

  3. 03

    Relationships between pieces of information are stored, so answers keep their context

  4. 04

    Data security designed in as a product priority from the start

  5. 05

    The site demonstrates rather than describes: a RAG assistant over MAIstr's own knowledge is there to try

  6. 06

    Qdrant for vector search and PostgreSQL for structured data

Results

  • Know-how stays with the company even after an experienced employee leaves

  • Onboarding a new person relies on answers instead of hunting through documents

  • Answers hold context across sources rather than matching text

  • Visitors try RAG for themselves before speaking to anyone

  • The product is running a pilot with its first customers

Development timeline

From analysis to deployment - how the project evolved

  1. ongoing

    Validating the problem

    Problem identified in the family business, then interviews confirming other companies face it too

  2. iterative

    Knowledge base prototype

    AI processing of inputs, design of relationship storage, vector search in Qdrant

  3. iterative

    Application and agents

    React and Vite interface, Python backend, specialised agents and training workflow

  4. iterative

    Website and product launch

    Next.js and Tailwind marketing site with its own RAG assistant, measured with Google Analytics and Clarity

  5. current

    Pilot operation

    Taking on first customers, feedback from real usage and reprioritising accordingly

Have a similar project?

I'll be happy to help you build your application. Feel free to contact me for a free consultation.

Contact me