Loading the globe…
Opt in the Browser

Solve optimization models,
right in your browser.

Model and solve LP & MIP problems with GLPK, HiGHS, CBC and SCIP. Real solvers compiled to WebAssembly.

Open editor

Browser workspace · Local folder · Google Drive Sync
MathProg reads CSV data · Solving runs locally

From formulation to solution, without setup.

Opt in the Browser is a zero-install workspace for linear programming and mixed-integer programming. It provides a modern, cross-platform editor with four solver engines and runs optimization jobs without sending them to a server.

Write GNU MathProg / GMPL or open an LP or MPS model, inspect the solver log, compare engines and keep the project in browser memory, a local folder or optional Google Drive storage.

One workspace. Four real solvers.

Native projects compiled to WebAssembly, each isolated in a Web Worker.

GLPK
MathProg / GMPL · LP · MPS
Native MathProg engine, model conversion and report generation.
HiGHS
LP · MIP · MPS
High-performance linear and mixed-integer optimization with strong presolve.
CBC
LP · MIP · MPS
The established COIN-OR branch-and-cut engine for mixed-integer models.
SCIP
LP · MIP · MPS
An advanced solver for harder discrete and constraint-integer models.

A shorter path from formulation to feedback.

  1. Open a real model

    Start with the bundled diet, Sudoku, travelling-salesperson and introductory examples, or open your own project without creating an account.

  2. Edit model and data together

    Keep .mod, .dat, .csv, .lp and .mps files in one workspace. Monaco provides syntax-aware editing while the explorer keeps multi-file models together.

  3. Run the solver off the UI thread

    Choose GLPK, HiGHS, CBC or SCIP. The WebAssembly solver runs in a dedicated Web Worker and streams its output into the console.

Questions worth answering before the first solve.

No benchmark theatre and no claim that a browser replaces every native workflow.

Is Opt in the Browser free to use?
Yes. There is no account requirement or per-solve fee for browser and local-folder workspaces. Google Drive sign-in is optional.
Is it suitable for a MathProg course?
For many classroom workflows, yes: it combines .mod and .dat files, an editor, examples, solver controls and a console in a cross-platform browser app. Courses that need desktop-only GLPK features or command-line options not exposed by the web interface should still use native tooling.
Do models leave the browser when they are solved?
No. Solver execution stays on the device. Browser workspaces need no server, local folders are read directly when the browser supports them, and optional Google Drive files move directly between the browser and Google APIs.
Can a MathProg model run with HiGHS, CBC or SCIP?
Yes. GLPK converts the model to MPS, the selected solver computes a primal solution, and GLPK reconstructs the model report. Solver-specific dual values and basis statuses are not portable through that path.
What model size can the online solvers handle?
The practical limit depends on the device, browser and solver. This workspace is best for teaching, prototypes and moderate experiments; large or production-critical models should still be verified with a native solver.