icom/ framework

Evidence before claims.

Regression tests and repeatable measurements are part of the source.

Run the checks

example / terminal
composer validate --strict
composer audit
composer test
find src app routes public -name '*.php' -print0 | xargs -0 -n1 php -l
php bin/icom benchmark 10000

What the suite covers

Route methods and parameter encoding, dependency injection and circular references, scoped isolation and reset, middleware ordering, safe error responses, input parsing and body limits, CSRF expiry and tampering, validation edge cases, PDO injection resistance and rollback, cache integrity/TTL, rate-limit locking and encryption tampering. The release process also checks the live HTTPS site and desktop/mobile playground interactions.

Interpret the microbenchmark

The CLI benchmark warms one application and dispatches an in-memory GET route repeatedly. It reports elapsed time, p50/p95 latency, requests per second and peak PHP allocation. It excludes network, TLS, database, Composer bootstrap, real client concurrency and the production middleware stack. It is not a public throughput figure and cannot establish superiority over Laravel or Symfony.

A fair comparison

Use the same machine, PHP build, OPcache settings, server/runtime, workload, payload and middleware for each framework. Measure cold starts, warmed JSON routes, database reads/writes and realistic error/auth paths. Report p50/p95/p99, CPU, memory, error rate and concurrency; publish application source and command lines. No comparative benchmark has been run for this preview.

Current reference points

The development prompt was checked against Laravel 13 (PHP 8.3+) and Symfony 8.1 (PHP 8.4+). ICOM takes these as design references, not evidence of feature parity. See the official release notes and worker-state guidance below.

Laravel 13 releases ↗

Symfony 8.1 release ↗

FrankenPHP worker guidance ↗

Developer preview — benchmark claims require reproducible evidence.

Explore the architecture