The KhaBot mission, in five notes, part 1 of 5
Why KhaBot exists
The mission: close the last step.
I have been building software for more than fifteen years. I ran my own software company. I use AI every day, for almost everything.
And for the last few years I have been losing a fight that I should be winning.
The fight is simple to describe. I try to keep track of my accounts, my payments, my data and my security, across every service I depend on. Every month there are more of them. Every month it gets harder.
I am not alone. Every power user I know is in the same fight. Every business I know is in it too.
What makes me angry is that it should be easy by now.
Everything already exists
Look at what we have today.
- Most paid services have an open-source alternative.
- Cloud machines, storage and GPUs can be rented by the second.
- AI can rebuild most of the small products we pay for every month.
- AI can connect all of these things together and run them for you.
Put those four facts next to each other. A person or a company should be able to run most of their digital life for very little money, in private, on their own terms.
We can't. Not in practice.
Each of those things sits behind its own account, its own minimum payment, its own monthly bill, its own dashboard and its own key. Each one is small. Together they are a wall.
And AI made the wall higher. There are more services than ever, and each one wants a copy of your data. Many of them are a thin layer on top of someone else's model, and that thin layer sees everything you send.
One step away
Here is the part that hurts.
I keep meeting people who are one step away from something that would save them money, or protect their data, or give them back control. One subscription away. One tool away. One missed update away. One missed tweet away.
An electrical engineer pays far too much for speech, because he has never heard of a much cheaper service that is good enough for his project. It is not his field.
A young developer builds a whole app with AI and then stops, because he does not know where to rent a server or how to connect a domain.
And then there is me. Following new models is my job. I know when a cheaper and better one is released. I still have old projects calling the old one, because updating them costs me more time than it saves. So I keep paying.
If I don't take that step, who will?
Nobody is missing the technology. They are missing the last step.
I could write articles about all of this. But every one of them would have to begin the same way: subscribe to these twenty services, bring their keys, and do these ten things by hand.
Who would read that? Who would be able to follow it?
So the first step could not be an article.
Step one: one key
KhaBot started as something I needed for my own company. I began at the start of 2025, after dozens of attempts that failed and taught me what does not work. For two years the only users were me, a few friends and my employees. I rebuilt it from scratch more than once.
The idea fits in one sentence. One account, one key and one balance for every service you need.
AI models. Web search. Speech. Images. Storage. Linux machines. More are planned.
Each request costs what the provider charges. KhaBot's fee is taken once, when you add credit. When you use nothing, you pay nothing. Every key can have its own limit. And when you want to leave, you leave, because the APIs are the standard ones.
KhaBot does not own most of what sits behind that key. It routes. I say this openly, because the work was never about owning machines. It was about removing the step.
Step two: guidance
Once everything is behind one key, a guide can begin with one line.
Take a KhaBot key.
Then an AI agent can do the rest: read the guide, do the work, and tell you what it cost. The result is yours. The scripts sit on your machine, and you can read them, change them, or ask someone you trust to check them.
That is the second step, and it is only beginning. First your money. Then your data. Then your attention.
The other gap
There is a second gap, and it has nothing to do with services. It is between people and the AI itself.
Some people only chat with AI. Some let an agent work for them. A few run whole teams of agents. Most people cannot reach that last level. They are smart enough. The door was built for experts, or for the companies that sell the models.
So KhaBot is also an app. It runs on your own machines, and it is a place where you and your agents work together. I will come back to it in another note.
The rule underneath
I do not think everyone should run their own server. I do not think everyone should cancel their subscriptions, or work with ten agents at once.
I think everyone should be able to.
There is a difference between being and being able to be. That difference is the whole project. Nothing in KhaBot is forced. Use one service or all of them. Stay or leave. Work the way you already work, and take the next step only on the day you need it.
For every decision I ask myself one question. Does this make a better world? Does it make AI cheaper, closer, and open to more people, sooner?
We do not know what people can do when they are given the right tools.
I want to find out.