The Problem
`total === expected` looks harmless until binary rounding turns equal numbers into near-equal failures or false matches.
Safe floating point comparison validator
Catch float precision bugs before they corrupt calculations in production.
float-comparison-checker scans your codebase for unsafe equality checks, recommends epsilon-safe replacements, and enforces precision discipline directly in CI.
The Solution
AST-based analyzers identify risky comparisons and output safe Math.abs(a - b) < EPSILON or math.isclose(...) fixes.
Why Teams Pay
Fintech and data-heavy backends avoid flaky tests, reconciliation bugs, and costly postmortems with a lightweight guardrail.
What you get on day one
- CLI scanner across JavaScript, TypeScript, and Python projects.
- Web API to analyze snippets in pull request checks or internal code quality portals.
- Risk scoring dashboard that highlights highest-impact precision hotspots first.
- CI docs and commands to enforce precision checks as a required merge gate.
$8/month
Protect CI pipelines against silent float precision regressions.
- JS, TS, and Python analyzers with actionable suggestions
- CI-friendly non-zero exit codes to block risky merges
- Web dashboard with risk score and remediation snippets
- Stripe hosted checkout with cookie-based access unlock
Buy Now
Cancel anytime. Teams can scale to yearly plans through support after onboarding.
FAQ
What exactly does the scanner flag?
It detects direct float equality and inequality checks like ==, ===, !=, and !== when values look derived from decimals or math operations.
Does it work in CI pipelines?
Yes. The CLI returns a non-zero exit code by default when issues are found, so GitHub Actions, CircleCI, and Buildkite can fail fast.
Will it flood us with false positives?
The analyzer uses AST context and float heuristics. You can scope extensions, adjust epsilon recommendations, and suppress expected exceptions in code review.
Who should deploy this first?
Teams with monetary calculations, experiment analytics, geospatial math, and scientific transforms where one rounding mismatch can produce bad decisions.