Free sees what's wrong.
Pro fixes it.
8 capabilities are free forever — the layer that finds problems: the app map, impact checks, security scans, graded audits. 8 more unlock with Pro — the layer that does the work: plans, automatic tests, launch pipelines, and the full build-it-right rulebook.
Forever free — personal use, all setup skills, 5 reviews/day. See every problem — CVEs, coverage gaps, architecture drift, the blast radius of any change — before you pay for anything.
Code Knowledge Graph
See how your code connects — callers, callees, blast radius before any change.
- Live graph updates on every save
- Impact analysis before refactors
- Dependency chains across files
/composure:build-graphPR Review with Blast Radius
Graph-aware code review that catches what eyeballs miss.
- Impact-aware review (which files break if this lands)
- Quality and security checks combined
- Local diff review or full PR
/composure:critique-local/composure:critiqueSecurity Scanning (OWASP)
Find CVEs, leaked secrets, and missing headers — before they ship.
- Dependency CVE audit
- HTTP security header analysis
- Pre-write secret-leak and insecure-pattern guards
/sentinel:assess/sentinel:audit-deps/sentinel:headersTest Framework Calibration
Detect your test setup and surface coverage gaps — automatic.
- Detects Jest/Vitest/Playwright/Pytest setups
- Runs full suite or just changed files
- Coverage-gap nudges on save
/testbench:calibrate/testbench:runDeployment Readiness
Production preflight checks — env, health, security, dep health.
- Env-var and health-check audits
- Validate existing CI workflows
- Dep health (CVEs, outdated, unsafe versions)
/shipyard:ci-validate/shipyard:deps-check/shipyard:preflightArchitecture Guides & Codebase Audits
Per-stack architecture playbooks and full-codebase health audits with letter grades.
- Stack-aware architecture guides
- Audit reports with scored grades
- Changelog generation from git history
- Pre-write code enforcement (no-bandaids, framework drift)
/composure:app-architecture/composure:audit/composure:changelogCommit & Session Hygiene
Conventional commits, session rollups, MCP install — automated workflow plumbing.
- Commit messages with task queue hygiene
- Session summary at end of each session
- One-command MCP server install
/composure:commit/composure:mcp-install/composure:session-summaryControl Panel
Mission control for the fleet — sessions, campaigns, graph, and plans in one pane of glass.
- Live sessions + agent fleet map
- Campaigns board — phase/stage with live timer (Pro)
- Code graph + blueprints kanban, local-only
Free plan limits
5 PR reviews / day (metered) · Personal use only. Everything else above has no ceiling.
See the full Free vs Pro breakdownNo card required.
- sessions in the archive
- 820
- sessions in the archive
- projects
- 70
- projects
- active build time
- 850h
- active build time
- tool calls
- 93,374
- tool calls
- commits
- 6,040
- commits
- blueprint runs
- 105
- blueprint runs
Counted by a census script over the session records the system writes about itself (2026-03-25 → 2026-08-15) — every definition and caveat ships inside the data. Hours and commits cover the 42% of sessions carrying per-session stats, so both are floors. The same archive every review on this page is mined from.
Your AI's memory is searchable.
Its reasoning is visible.
Cortex remembers across sessions — every decision, the thinking behind it, and the graph that links it all. The Control Panel is where you read it back: sessions, campaigns, memory, and the code graph on one screen. Staged demo below — click through it; yours renders live.
▸aurora-crmGo deeper
The full story behind what Pro unlocks.
Pro vs other coding agents
How Composure enhances Claude Code and Codex — a memory graph, code map, blueprints, and enforcement they don't ship natively.
Cortex — your AI's memory
Searchable memory, the graph that links it, and the reasoning behind every decision — all inspectable.
The skill catalog
Every Pro skill, by purpose — and the quota story: Free is metered, Pro is unlimited.
Templates
Production-grade starters — what you get, and the hours they save versus building from scratch.
App architecture, in practice
The one-hour AI app looks done and is quietly broken — slow SSR, no cache, secrets in the bundle, no RLS. See what building it right the first time looks like.
What it delivers
Measured runs, not promises — 86% cheaper, 30× fewer tokens, a full dashboard built in 30 minutes.
Who it's for
Exploring what's available
Install free and get the full diagnostic layer — the code knowledge graph, security scanning, architecture audits with letter grades, deployment preflight, and 5 PR reviews a day. Personal use, no card, no ceiling on the diagnostics.
Ready to unlock the action layer
Pro is for shipping: blueprints before complex work, generated tests and CI/CD, integration scaffolds, deep security scans, the 101-pattern architecture catalog — and everything in Free goes unmetered, with unlimited PR reviews and commercial use.
Building for clients or a team
Commercial use on any plugin tier requires Pro or above. Agencies and multi-seat teams → see Team on the pricing page; commercial-license details are at /legal/plugin-license.
See your own blast radius — before you pay for anything.
Install the diagnostic layer on the repo you already have. No card required. Personal use, nothing gated.
$pnpm dlx create-composureYou'll get a code knowledge graph, a blast-radius map of any change, a security sweep, and an architecture audit with letter grades. That's the whole diagnostic layer, free, running on your own code.
Before you decide.
The five questions that come up before installing.
$39/mo. Or $24/mo annual founding rate.
Start free, or go Pro — monthly or annual, cancel anytime. Review-only tools stop at telling you what's wrong; Pro adds 22 multi-tenant DB patterns, blueprints, and the full architecture catalog — the part that fixes it.