Technology Keeps Changing. So Do We.
TL;DR mini Labs records how Minisoft prepares, tests, and applies AI and business software in real work. The focus is not technology change by itself, but how that change becomes practical products and workflows.
Two years of AI preparation, and a record of change that continues
Over the past two years, we have been preparing for LLM-based AI technology.
At first, the work started with building simple chatbots. But over time, the questions changed.
- Can AI understand documents?
- Can AI understand images?
- Can enterprise data be connected safely?
- Can multiple AI Agents work together?
To answer these questions, we tested many technologies and carried out multiple PoCs and pilot projects.
Some experiments worked better than expected. Some were harder than expected. But every step helped us grow.
Technology trends change faster than expected
The pace of technology change is fast. Something that felt cutting edge six months ago can quickly make room for a new approach.
After the rise of RAG, multimodal LLMs became more practical, and MCP is becoming a new standard for AI Agents.
Our internal technology stack is changing just as quickly.
Our company and solutions are evolving together
We continue to upgrade our solutions in line with technology trends.
Starting from simple chatbots, we have expanded into RAG-based knowledge search, Vision AI, multimodal document understanding, MCP-based AI Agents, Enterprise Security, and Mini App Platform.
The solutions we build are less like finished products and more like platforms that continue to evolve.
What we want to record
MiniSoft Labs is not a place for showing only finished answers.
It is a place to record which technologies we chose, what we tested, what worked, and what was difficult.
Technology keeps changing. So do we.
MiniSoft Labs will record that journey of change.
Subscribe to new posts
Follow new posts through the RSS or Atom feed.
Reactions and comments
Sign in with GitHub to comment and leave a reaction (like). Comments and GitHub usernames are public; editing, deletion, and reporting are handled through GitHub Discussions.