I built KalaBodha because I wanted an AI assistant I could trust.
Not trust as a feeling. Trust as a verifiable property. Trust I could inspect, audit, and prove.
I have spent 22 years as an electrical engineer designing safety-critical systems — power generation, substations, distribution networks. In that world, a system earns trust through transparency: every wire is traceable, every protection relay is testable, every fault is logged. You do not hope a power grid is safe. You verify it.
When I started using AI assistants — and they are genuinely impressive tools — I noticed something that my engineering training would not let me ignore. I could not verify anything. I could not trace how an answer was produced. I could not audit what the system did with my data. I could not prove that my contacts, my decisions, my communication patterns were not being used to train the next model version.
I did not set out to criticize these tools. I set out to build one that meets the standard I was trained to expect.
KalaBodha is an AI assistant powered by a knowledge graph you own.
It learns your contacts, your relationships, your communication patterns, your decisions, and your preferences — structurally, in a graph that grows with every interaction. Not in model weights that are frozen at training time and controlled by someone else.
When KalaBodha tells you something from your knowledge graph, it shows you the source. When it generates content using the language model, it tells you that too. You always know what is verified and what is assisted.
Your data is encrypted with keys only you hold. The system runs on your hardware. It works offline. The code is open source (AGPL-3.0). You can inspect every line.
This is what I mean by trustworthy AI. Not a promise. An architecture you can verify.
Where this is headed.
I believe we are at the beginning of a shift. The current era of AI is model-centric — intelligence lives in weights, controlled by the companies that train them. The next era will be knowledge-centric — intelligence lives in structured, user-owned knowledge that is auditable, versioned, and permanent.
The language model is an extraordinary tool for communication, creativity, and planning. It should be used as exactly that — a tool. The brain should be something you own. Something that grows with you. Something you can audit, export, and trust.
KalaBodha is the first product on this path. The personal assistant is the entry point. The architecture is the contribution.
I am building this as a solo founder, in the open, with an open call for collaborators. If this resonates — if you believe AI should be trustworthy, verifiable, and genuinely useful — I would be glad to build alongside you.
"Here is my answer. Here is why. Here is the proof. Here are my limits."