A collection of seventeen browser-native, fully accessible interactive tools for an introductory numerical analysis course at the junior level, suitable for mathematics, engineering, computer science, and computational science majors. All tools run entirely client-side with WCAG 2.1 AA conformance — keyboard navigation, screen-reader-accessible mathematics, high-contrast modes, and visible focus indicators throughout. The collection is organized to support direct comparison of methods on shared problems, an approach that is awkward to set up in classroom MATLAB or Python environments and is rarely available in a single integrated tool. Three root-finders sit side by side: bisection (with interval zoom and iteration table) demonstrates guaranteed but linear convergence; Newton's method, animated through tangent-line iterates, exhibits quadratic convergence and the specific failure modes (cycling, divergence, attraction to an unintended root) that initial-guess sensitivity produces; fixed-point iteration is presented through a cobweb visualization with live Lyapunov-exponent calculation, classifying stability and exposing the period-doubling route to chaos as a numerical-analysis phenomenon rather than a separate dynamical-systems topic. Polynomial interpolation supports discussions of the Runge phenomenon and node placement, and is paired with two least-squares regression tools — polynomial and rational — for direct comparison of regression families on identical data. A finite-difference stencil generator builds forward, backward, and central formulas on demand with truncation-error estimates. A numerical-integration tool covers the full progression from left, right, midpoint, trapezoidal, Simpson, and Gauss-Legendre quadrature through adaptive Simpson and Romberg extrapolation, with convergence-rate comparison and tolerance-driven subdivision. The collection is cross-listed with the author's linear-algebra portfolio, exposing the algorithmic core of numerical methods (LU with partial and scaled-partial pivoting, QR, Cholesky, SVD, Gram-Schmidt, and three parallel least-squares solvers with explicit annotation of condition number and stability tradeoffs) as the foundation for solving linear systems, least-squares problems, and eigenvalue problems.
