Every hard problem
has a best answer.
The problem
You schedule shifts by hand. You split costs by feel. You pick projects in meetings where the loudest voice wins. You stare at a research conjecture and wonder if your approach even works.
Then you ask an AI chatbot and it gives you a confident-sounding answer that might be wrong. You have no way to know. It sounds just as confident either way.
Every one of these problems has a mathematically best answer. You just never had a way to find it.
The hidden structure
The scheduling problem on your whiteboard? That's a constraint satisfaction problem. The rent split argument? That's a fair division problem with a provably envy-free solution. The project selection meeting? That's a knapsack optimization. The eigenvalue computation in your research? That's a well-studied numerical linear algebra problem.
These aren't metaphors. These problems — from everyday decisions to open research questions — have been studied for decades. Algorithms exist that find the best answer and prove it's the best. They run in seconds.
The gap was never the math. It was access. The tools existed for PhDs and operations researchers. Not for everyone else — the business owner, the student, the engineer, the curious researcher.
Two ideas, 180 years apart
In 1837, Charles Babbage designed the Analytical Engine — the first general-purpose computer. Not a calculator for one kind of problem. A machine that could be programmed to solve any computation. The idea was radical: one machine, any problem.
Over a century later, in 1945, George Polya published How to Solve It — the first general-purpose problem-solving algorithm. Not a formula for one kind of equation. A method that could approach any problem. The same idea, for human thinking: one method, any problem.
Babbage's vision inspired us — one engine for any problem. Polya's method drives us. polya. is built on his algorithm: understand, plan, execute, verify. Applied to every problem you face — personal, professional, or pure research.
The method
Polya's universal algorithm has four phases. We modernized them with real solvers, real verification, and plain language input — but the structure is his.
Understand
What are you actually trying to solve? What are the constraints? What does 'best' mean here?
Plan
What kind of problem is this? What tools exist to solve it? Have similar problems been solved before?
Execute
Run the solver. Find the answer. Not a suggestion — the mathematically best result given your constraints.
Verify
Check the answer independently. A different algorithm, a different approach. If it holds, you have proof.
Polya wrote this for mathematics students with pencil and paper. We built it for everyone with a browser. You describe a problem in plain language. The system understands it, plans the approach, solves it, and proves the answer is correct.
Built on open source
At the core of polya. is uber-polya — an open-source coding agent skill that implements Polya's method as a working algorithm. It classifies problems, selects solvers, generates code, executes it, and verifies the result.
This app is a consumer interface to that skill. But the skill itself is free and runs anywhere. Load it into Claude Code, OpenAI Codex, Cursor, GitHub Copilot, or any compatible platform — type /uber-polya and solve.
We built the app for people who don't live in a terminal. But the engine is the same. Same method, same solvers, same verification. Whether you use the app or the skill, Polya's algorithm is what drives the answer.
Try the skill directly
# Install git clone https://github.com/agtm1199/uber-polya.git cd uber-polya bash install.sh # Then solve /uber-polya Schedule 12 nurses across 3 shifts
What we believe
Answers deserve proof, not confidence
A confident answer isn't a correct answer. When an AI chatbot says 'here's a solution,' you have no idea if it's right. When polya. says 'here's the optimal answer,' every constraint has been checked and the proof is included.
Math should be invisible
You shouldn't need to know what an integer linear program is to get the best shift schedule. The math is there — running real solvers, real algorithms — but you never see it unless you want to.
Honesty over confidence
When a problem can't be solved, we tell you. When constraints conflict, we show you which ones. When the answer is approximate rather than optimal, we say so. An AI that never says 'I don't know' is an AI you can't trust.
Answers should be checkable
Every solution includes the code that produced it. Every answer is verified by a different algorithm than the one that found it. Transparency isn't a feature — it's a requirement.
Everyone deserves optimal
Fortune 500 companies have operations research teams. Research labs have PhD students. Everyone else has spreadsheets and gut feel. The best answer shouldn't depend on who you are or what you can afford.
Why “polya.”
Babbage showed us the vision — one machine for any problem. Polya gave us the method that actually runs inside it. The name honors the method, because that's what drives every solution.
The period is intentional. It's a full stop. The answer isn't a conversation. It isn't a suggestion. It's the answer, period.
Don't guess. Solve.
Describe your problem. Get the best answer. With proof.
Get Started Free →