For courses and prototyping: the circuit, simulated; the algebra, computed.
The matrix side - Pauli algebra, tensor products, unitaries, eigenstructure - evaluates interactively on any plan. Circuits run through an ideal statevector simulator as background jobs, and come back as a measurement histogram.
It computes from your model, not from your files: Qurak solves what you write down, and does not read spreadsheets or datasets.
This is education and prototyping, said plainly: an ideal statevector simulation with a qubit ceiling per plan - not a hardware backend, not noise models, not a research cluster. If your work needs thirty qubits, we are not your tool yet, and this page will not pretend otherwise.
Every example here was run by the engine that will run yours.
| You write | It answers |
|---|---|
| KroneckerProduct[PauliMatrix[3], PauliMatrix[3]] | {{1, 0, 0, 0}, {0, -1, 0, 0}, {0, 0, -1, 0}, {0, 0, 0, 1}} |
| Eigenvalues[PauliMatrix[2]] | {-1, 1} |
| MatrixExp[{{0, -t}, {t, 0}}] | {{Cos[t], -Sin[t]}, {Sin[t], Cos[t]}} |
Plot[Exp[-x/5] Sin[2 x], {x, 0, 10}] · Plots render as SVG, in the workbench and over the API.Qurak is an MCP server, so Claude, Cursor, VS Code and any other MCP client can hand it a calculation and get a real answer back instead of a plausible one. There is an HTTP API for everything else. Both are open on the free plan.
Circuit jobs run to 22 qubits on the paid plan, more by conversation - the ceiling is the honest upgrade path.