Zeretis is a symbolic computation engine for the problems that come up constantly in engineering work — ODEs, systems of equations, matrix operations, calculus. CPU-only, deterministic, and accessible via browser or REST API.
When you're in flow you want to enter expressions as fast as possible. Zeretis accepts standard engineering notation, full LaTeX, and plain English descriptions — use whichever is fastest in the moment. All three are interchangeable within the same session.
Zeretis covers the core symbolic computation tasks that come up repeatedly in mechanical, electrical, civil, and software engineering contexts.
Derivatives (including partial), indefinite and definite integrals, limits, Taylor and Maclaurin series, L'Hôpital's rule. All returned as exact symbolic expressions.
First and second order ODEs, systems of ODEs, initial value problems. Solved symbolically using integrating factors, variation of parameters, and characteristic equations where closed forms exist.
Matrix operations, determinants, inverse, eigenvalues and eigenvectors, row reduction, rank, null space, LU decomposition. Exact rational results — no floating-point error.
Linear and nonlinear systems, solved symbolically. Useful for circuit analysis, structural equilibrium, thermodynamic state equations, and control system design.
Laplace transforms and inverse Laplace transforms for control systems and signal processing. Exact symbolic results directly applicable to transfer function analysis.
Integrate the full solver into your own tools. Send an expression via HTTP POST, receive exact symbolic results as structured JSON — including the step-by-step breakdown.
Numerical solvers like MATLAB's ode45 or numpy give you floating-point approximations. For many engineering tasks that's fine — but there are situations where it isn't. If you need a result that will be used in further symbolic manipulation, an approximation breaks the chain. If you need an auditable, reproducible result, a floating-point number that varies with step size isn't it.
Symbolic results are also inherently more general. The closed-form solution to a differential equation tells you how the system behaves for all initial conditions — not just the one you happened to plug in. That understanding is often more valuable than a numerical trajectory.
Zeretis returns exact results — fractions stay as fractions, surds stay as surds, constants like π and e are preserved symbolically. This means you can use the output directly in further derivations, substitute values in later, or verify that the form matches what you expected from theory.
And because symbolic computation doesn't use stochastic processes or neural networks, results are completely deterministic. The same expression run twice, on different machines, a week apart, gives the same result. That reproducibility matters for engineering documentation, peer review, and audit trails.
Symbolic computation is algorithmically efficient — it doesn't require neural network inference, GPU memory, or massive parallel compute. Zeretis runs on standard server hardware, which matters if you're self-hosting the API or thinking about integration at scale.
Zeretis uses similar symbolic computation techniques to SymPy and Mathematica — the underlying math is the same class of algorithm. The difference is accessibility and integration: Zeretis is available via browser with no installation, and via a simple REST API, making it easy to embed in tools, workflows, and scripts without managing a Python environment or a Wolfram license. For exploratory work or API integration it's significantly more convenient; for large-scale or highly customised computation Mathematica still has a deeper feature set.
Yes — that's exactly what the REST API is for. POST an expression to our endpoint, receive structured JSON back including the symbolic result, LaTeX representation, and step-by-step breakdown. See the enterprise page for API documentation and pricing.
The solver tells you — it doesn't approximate or return a numerical result silently. If a closed-form symbolic solution isn't possible for the given expression, you'll get an explicit message to that effect rather than a misleading approximate answer.
The standard API plan includes 10,000 calls per month. Beyond that, additional calls are billed per use. Enterprise plans with higher volumes and custom SLAs are available — see the enterprise page or contact us directly.
Try the web interface free. API access from $10/month including 10,000 calls.