Frequently asked questions

Cartara is a local-first developer platform that measures engineering code comprehension in AI-assisted teams. It turns every AI coding session into a living knowledge graph — an evidence-based picture of what your team actually understands about the code it ships, not just how fast it ships.

When an AI agent does the reasoning, engineers can ship correct code for months without ever building the mental models the work depends on. Velocity looks great while a silent gap opens between what was shipped and what is understood — and that gap surfaces at the worst times: incidents, security reviews, migrations. Cartara measures that gap and helps close it.

Cartara is built for AI-native engineering teams and their leaders — anyone who wants delivery speed and durable human understanding of the code being shipped. There's also Cartara for Education, a project-based version for students, teachers, and schools.

Delivery metrics tell you how fast code ships; they say nothing about whether your team understands it. Cartara treats understanding as a first-class metric alongside delivery — a per-concept mastery map grounded in evidence from your own repository, not agent self-report.

Cartara doesn't monitor keystrokes or activity. It grounds understanding in real work: plain-language reviews generated from your actual diffs, plus fast three-minute knowledge checks tied to the code just shipped. The result is a per-concept mastery map at the team level — where you're strong, where you're fragile — with no surveillance and nothing to game through agent self-report.

A comprehension check is a short, roughly three-minute set of evidence-based questions tied to code your team just shipped. Because each question is grounded in a real diff, mastery reflects understanding of the actual work — not trivia, and nothing that can be gamed by agent self-report.

The knowledge map is a living, per-concept picture of what each engineer and team genuinely understands, built up from the reviews and comprehension checks tied to real diffs. It shows where your team is strong and where it's fragile, and it updates as new work ships.

Cartara is local-first. The collector runs on your machine and applies your privacy policy before anything is uploaded: denied paths like .env stay local, secrets are redacted, and you choose metadata, sanitized, or full detail per project.

You set the detail level per project — metadata only, sanitized, or full detail. The collector applies your privacy policy before anything is uploaded: denied paths like .env stay local and secrets are redacted, so you decide exactly how much of each repository Cartara ever sees.

Cartara works alongside the AI coding tools engineers already use — including Claude Code, Cursor, GitHub Copilot, and Codex — because it learns from the diffs those sessions produce rather than hooking into any single vendor.

Yes. Cartara connects to GitHub through a GitHub App so it can analyze the pull requests and diffs your team ships, then ground reviews and comprehension checks in that real work.

On a running Cartara instance you connect your GitHub account and keep using the AI coding tools you already have — Cartara learns from the diffs those sessions produce. See the Quick Start guide for step-by-step setup once you have access.

Cartara is rolling out to teams from the waitlist. Join the waitlist to request access — we'll reach out as spots open.

Public pricing isn't announced yet. Join the waitlist and we'll share plan details with you directly as they're finalized.

Yes — Cartara for Education is a project-based version for students, teachers, and schools: learn software engineering by building real software, with evidence-linked comprehension instead of AI-generated homework.