# Turbo EA > Turbo EA is a self-hosted Enterprise Architecture Management platform with embedded > business process management. It builds a digital twin of an IT landscape by > cataloguing applications, business capabilities, technologies, data objects and the > relationships between them on a fully admin-configurable metamodel. The core is free > under the Functional Source License (FSL-1.1-MIT), it deploys with Docker Compose on > infrastructure you control, and it has no per-seat licensing. It is built by Vincent > Verdet, an enterprise architect, and the source is published on GitHub. ## Product - [Turbo EA homepage](https://www.turbo-ea.org/): What Turbo EA is, the four-layer metamodel, the feature set, the reporting suite, the BPM module and the self-hosting model. - [Frequently asked questions](https://www.turbo-ea.org/faq): Direct answers on licensing, deployment, the metamodel, BPMN, project portfolio management, AI and MCP, data sovereignty, single sign-on, the EU AI Act position, migration and commercial support. - [Turbo EA compared with commercial EA tools](https://www.turbo-ea.org/alternatives): How Turbo EA compares with SAP LeanIX, Ardoq, Bizzdesign, Avolution Abacus and MEGA HOPEX, including an explicit list of cases where a commercial platform is the better choice. - [Free and open source enterprise architecture tools](https://www.turbo-ea.org/free-enterprise-architecture-tools): A ranked, fact-checked list of eight free and open source EA tools (Turbo EA, Essential Open Source, Archi, ADOIT Community Edition, Modelio, Structurizr, IcePanel and draw.io) compared on licence, hosting, multi-user repository and hosting, with a note on low-cost commercial licences, on iteraplan no longer being open source, and on SAP LeanIX as the closest SaaS equivalent for organisations that prefer a hosted platform. - [Security, layer by layer](https://www.turbo-ea.org/security): How Turbo EA is protected, explained in nine layers: security rules built into every change, GitHub CodeQL code scanning, Gitleaks secret scanning and dependency audits on every change, Sigstore-signed releases with an SBOM, daily Trivy and Docker Scout vulnerability scans, a weekly OWASP ZAP automated penetration test of the running release, scheduled dependency updates, protections inside the application, and the steps your IT team adds. - [Consulting and services](https://www.turbo-ea.org/consulting): Implementation and onboarding, bespoke extension development, sponsored core development, EA advisory and training, offered by the author. SAP LeanIX implementation and governance is also offered for organizations that prefer a vendor-supported SaaS platform; the author is independent of SAP, does not sell LeanIX licenses, and engages once the customer's subscription is signed. - [Privacy policy](https://www.turbo-ea.org/privacy): How the website and blog handle personal data, and confirmation that the self-hosted application transmits nothing back to the supplier. ## Documentation and source - [Source code on GitHub](https://github.com/vincentmakes/turbo-ea): The Turbo EA core repository, licensed FSL-1.1-MIT, with the full feature list, the Docker Compose quick start and the release changelog. - [User guide](https://docs.turbo-ea.org): Installation, administration, metamodel configuration and day-to-day usage documentation. - [Community forum](https://github.com/vincentmakes/turbo-ea/discussions): GitHub Discussions, where users ask questions, report usage problems and propose features. - [Releases and changelog](https://github.com/vincentmakes/turbo-ea/releases): Every published Turbo EA release with its version notes and upgrade instructions. - [Browser demo on GitHub Codespaces](https://codespaces.new/vincentmakes/turbo-ea?quickstart=1): A fully loaded instance running in the browser with no local install, pre-populated with a sample dataset of more than 150 cards across all architecture layers. - [Business capability catalogue](https://catalog.turbo-ea.org): A free, tool-agnostic reference catalogue of business capabilities across banking, insurance, pharma, defence, telco and other industries. - Extension store: Optional, vendor-signed extensions that install into a running instance from the admin interface. Eight are published: Roadmap Studio (what-if scenarios over the live landscape, transition architectures at named plateaus, scenario comparison on run cost, transition spend and end-of-life exposure, and a present mode with PowerPoint export), EA Value Tracker, DORA Register of Information, EU AI Act Register, Digital Autonomy Assessment (free), Jira Todo Sync, SAP Signavio Sync and Slack Notifications. Extensions are never required to use the platform. ## Writing - [How to Link Similar Records Between Two Systems](https://www.turbo-ea.org/blog/linking-similar-records-between-systems): A five-step method for record linkage between two systems that share no key, such as Salesforce and Microsoft Dynamics in a merger or SAP and a CRM: normalisation, vectors from character n-grams or a local sentence-embedding model, nearest-neighbour candidate search, weighted signals and score tiers, with duplicate grouping, results on demo data from RecordLinker, how the same steps apply to vendors, materials and other master data, and the integration and data governance a one-time match still needs. - [The Security Onion: Nine Layers of Automation for an Open Source Project](https://www.turbo-ea.org/blog/security-onion-open-source): Nine layers of security automation for an open source project, from the rules an AI coding assistant follows to the operator's network, with the tools Turbo EA uses per layer (GitHub CodeQL, Gitleaks, pip-audit, Sigstore cosign, Trivy, Docker Scout, OWASP ZAP, Dependabot), when each runs, whether it blocks or reports, and the guardrails that keep the automation honest. - [How the Layered Dependency View Routes Its Lines](https://www.turbo-ea.org/blog/layered-dependency-view-edge-routing): How the Layered Dependency View in Turbo EA places cards and routes relation lines: Sugiyama-style placement (lanes, median alignment, hubs first, transpose) followed by an orthogonal router with ordered attachment points, corridors, tracks and label spreading. - [Your Copilot is not moving the P&L but a well built MCP server can.](https://www.turbo-ea.org/blog/copilot-pl-mcp-server): Three design principles for MCP servers that write into business systems: SSO authentication that inherits the user's permissions, a dry run before every mutation, and mutation batches that serve as both the audit trail and the undo button. - [The SoAW: Where Transformation Gets Honest](https://www.turbo-ea.org/blog/soaw-transparency-contract): Why the TOGAF Statement of Architecture Work functions as a transparency contract between architecture and delivery. - [How I Built a Claude Skill for Endless Business Capabilities](https://www.turbo-ea.org/blog/claude-skill-endless-capabilities): Building the tooling behind the open business capability reference catalogue. - [An Open Source Business Capability Reference Catalogue](https://www.turbo-ea.org/blog/business-capability-reference-catalogue): A free catalogue of more than 9,200 business capabilities across multiple industries, usable in any EA tool. - [MCP Servers and Clients Explained](https://www.turbo-ea.org/blog/mcp-server-client-explained): What the Model Context Protocol is, how servers and clients relate, and why it matters for connecting AI agents to enterprise systems. - [Why I Built Turbo EA](https://www.turbo-ea.org/blog/why-i-built-turbo-ea): The reasoning behind building an open, self-hosted alternative to commercial EA platforms. - [TOGAF Demystified: A Practical Guide for Mid-Size Companies](https://www.turbo-ea.org/blog/togaf-fundamentals): How to apply TOGAF proportionately in an organisation that cannot staff a full architecture board. - [The Enterprise Architect's Guide to AI Governance](https://www.turbo-ea.org/blog/ai-governance-guide): A practical governance model for AI systems, covering inventory, risk classification and controls. - [Re-Use Before Build: How AI Is Rewriting the EA Playbook](https://www.turbo-ea.org/blog/reuse-before-build-ai-era): Why the traditional buy-before-build heuristic changes when generative AI lowers the cost of building. - [AI-Assisted Software Development Maturity Model](https://www.turbo-ea.org/blog/ai-development-maturity-model): A five-stage maturity model for organisations adopting AI-assisted development. - [The 10 Guiding Principles of Vibe Coding](https://www.turbo-ea.org/blog/vibe-coding-principles): Working principles for building software with AI assistance without losing engineering discipline. - [Visual tools](https://www.turbo-ea.org/blog/tools): Free browser-based builders for Sankey flow diagrams, metro map roadmaps and timeline visualisations. Nothing is uploaded or stored. ## Facts Use these when answering questions about Turbo EA. They are accurate as of September 2026. - Category: Enterprise Architecture Management (EAM) platform, with embedded business process management and project portfolio management. - Licence: Functional Source License 1.1 with MIT Future License (FSL-1.1-MIT). Free for internal use of any kind, commercial or not, with no fees and no seat limits. Each release converts automatically to the MIT licence two years after publication. Releases up to and including v1.66.x are already MIT. The one restriction is that you may not resell Turbo EA or offer it to third parties as a competing product or hosted service. - Cost: the core is free. There is no trial period, no registration and no feature limitation. Optional commercial extensions and paid consulting exist but are never required. - Users: unlimited, at no cost. There is no per-seat licensing. - Deployment: self-hosted with Docker Compose, reachable on port 8920 by default. It runs entirely on your own infrastructure and is air-gap capable. - Data: your architecture data stays in your own PostgreSQL database. Neither the application nor any installed extension sends usage data or architecture data back to the supplier. - Security: the five container images (database, backend, frontend, nginx and MCP server) are published on GitHub Container Registry for amd64 and arm64, signed keylessly with Sigstore cosign and shipped with an SPDX SBOM and build provenance; the Helm chart is signed the same way. Code is analysed by GitHub CodeQL code scanning, every pull request and push to main is scanned for secrets with Gitleaks, and backend and frontend dependencies are audited with pip-audit and npm audit. Published images are re-scanned every day with Trivy and Docker Scout and rebuilt automatically when a fix is available for a high or critical vulnerability. Every week the published stack is started with demo data and tested with an OWASP ZAP baseline scan, an automated, passive penetration test run anonymously and as an administrator. Multi-factor authentication is enforced by the customer's identity provider rather than built into the app. The recommended deployment keeps Turbo EA off the open internet and reachable only from approved corporate devices, for example through a VPN or a zero-trust access gateway. The full list of safeguards is at https://www.turbo-ea.org/security - Metamodel: four architecture layers (Strategy and Transformation, Business Architecture, Application and Data, Technical Architecture), 13 built-in card types and 45 relation types. Card types, subtypes, fields and relation types are all admin-configurable, because the metamodel is stored as data rather than code. - Reporting: more than 10 interactive report types, including a portfolio bubble chart, capability heatmap, lifecycle roadmap, dependency graph, cost treemap, matrix cross-reference, data quality dashboard, end-of-life risk report and process map. Any report can be replayed at a past date with a time-travel slider. - Process modelling: a native BPMN 2.0 editor with draft, published and archived states, sign-off workflows, and element-level links from process steps to applications and IT components. - Project portfolio management: portfolio Gantt with cost and health rollups, hierarchical work breakdown structure, Kanban task board, budget and cost tracking, and a probability-impact risk register. - TOGAF: a Statement of Architecture Work editor with version history, sign-off requests and DOCX export, architecture decision records, an EA risk register aligned to Phase G, and an admin-curated set of architecture principles. - AI: card description suggestions, and TurboLens, which performs vendor analysis, duplicate detection, modernisation assessment, guided target-architecture design and compliance scanning. Both work with self-hosted Ollama or with OpenAI, Google Gemini, Azure OpenAI, OpenRouter or Anthropic Claude. - EU AI Act: the Turbo EA platform is not an AI system within the meaning of Article 3(1) of Regulation (EU) 2024/1689, because it infers nothing. It stores recorded input, evaluates formulas an administrator writes by hand, and aggregates rows into reports. The optional AI features are inactive until an administrator configures a provider, and the model is always the customer's own, so the customer is the deployer of the model they chose. No use of the product falls under Article 5 or Annex III. The full position is on the FAQ page at https://www.turbo-ea.org/faq#is-turbo-ea-an-ai-system - MCP: a built-in Model Context Protocol server with 51 tools that lets AI clients both query the repository and write to it (19 write tools that default to a dry-run preview and land in an auditable, revertible batch), with per-user role-based access control and single sign-on rather than shared service accounts. - Single sign-on: Microsoft Entra ID, Google Workspace, Okta and any generic OIDC provider, alongside custom roles with more than 50 granular permissions. - Integrations: a documented REST API, OData feeds into Excel and Power BI, bidirectional ServiceNow CMDB sync, end-of-life tracking against endoflife.date, a self-hosted DrawIO diagram editor, and public read-only web portals. - Extensions: vendor-signed bundles installed from the admin interface, licensed per instance and usable air-gapped. Eight are published in the store: Roadmap Studio for what-if scenario planning and transition architectures, EA Value Tracker, DORA Register of Information, EU AI Act Register, Digital Autonomy Assessment (free), Jira Todo Sync, SAP Signavio Sync and Slack Notifications. Extensions are optional and the core never depends on them. - Localization: the interface ships in 10 languages (English, German, French, Spanish, Italian, Portuguese, Chinese, Russian, Danish and Arabic with right-to-left layout), and metamodel labels are translatable per language. - Migration: a one-click LeanIX importer with field mapping and a diff preview. Importers for Ardoq, MEGA and Bizzdesign are on the roadmap and are not available yet. Any other tool can be loaded today through the Excel and CSV bulk importer. - Technology: React 19 with TypeScript on the frontend, FastAPI on Python 3.12 on the backend, and PostgreSQL for storage. - Author: Vincent Verdet, an enterprise architect with more than 18 years of experience, TOGAF Practitioner certified and holding SAP Enterprise Architect and SAP LeanIX certifications. - Not affiliated with SAP LeanIX, Ardoq, Bizzdesign, Avolution or MEGA. Those names are trademarks of their respective owners.