Cranot (Dimitris)
Roam provides a local codebase intelligence layer for AI coding agents to map repo structure, catch structural risks, and generate audit evidence.
At a Glance
- Software engineering teams using AI agents
- Compliance-heavy organizations needing change evidence
- Open-source maintainers
AI Tools by Cranot (Dimitris)
(1)
roam-code
Local Codebase Graph for AI Agents
Discussions
No discussions yet
Be the first to start a discussion about Cranot (Dimitris)
Latest News
Roam releases benchmarks showing 80% reduction in agent input tokens
Roam announces pricing tiers and early access for Roam Review and Cloud
roam-code CLI version 1.0 released on PyPI
Roam cited in independent coverage of AI quality pressure (Faros AI report)
Products & Services
A free, local codebase intelligence engine that indexes repositories into SQLite to provide structural context for AI agents.
A hosted GitHub App that provides structural and algorithmic risk verdicts on every pull request.
A dashboard for tracking code health, debt, and danger-zone trends over time without receiving source code.
A paid one-shot engagement to audit a team's last 30 to 90 PRs and produce a structural review report.
Market Position
A structural intelligence layer that complements semantic reviewers (CodeRabbit, Greptile) by focusing on code connectivity and architectural drift.
Leadership
Founders
Dimitris Mitsos (Cranot)
Senior/Staff Engineer with expertise in developer tools, AI infrastructure, and code intelligence. Author of high-performance C++ and Python projects with over 4,600 GitHub stars, including mzip and tieredsort. Based in Athens, Greece.
Executive Team
Dimitris Mitsos
Founder & Lead Engineer
Solo creator of Roam and various low-level systems libraries.
Founding Story
Founded by Dimitris Mitsos to solve the 'verify' gap in AI-assisted coding, where agents generate code quickly but miss its structural consequences (callers, clones, hot paths).
Business Model
Revenue Model
Subscriptions for hosted PR reviews and dashboards; one-shot audit services.
Pricing Tiers
Apache 2.0 license, local engine, no telemetry.
3 repos, 5 active PR authors, 100 reviews/mo.
20 repos, 30 active PR authors, 900 reviews/mo.
100 repos, 100 active PR authors, 3,000 reviews/mo.
Dashboard for code health tracking.
One-shot audit of 30 PRs, written report, 50% credit toward Roam Review.
Target Markets
- Software engineering teams using AI agents
- Compliance-heavy organizations needing change evidence
- Open-source maintainers
- AI-assisted pull request review
- Codebase architectural mapping
- Blast-radius analysis for refactors
- SOC 2 / ISO 42001 change-management evidence