Open source · Self-hosted · Made in Germany

An open-source development and runtime platform built on a graph database.

Structr is where you build your application and where it runs. One Java server on top of Neo4j gives you a visual schema editor, an instant REST API, server-side rendered pages, BPMN workflows, object-level security and server-side JavaScript – develop and operate on the same platform, without a separate build or deployment stack. Self-hosted on your own infrastructure or in the Structr Cloud. In active development since 2010.

Try it nowDownload$docker compose up -d
  • Types
  • Data
  • Screens
  • Processes
  • Permissions
The Structr Pages editor with page tree, live preview and shared components
The Pages editor – page tree, live preview and shared components.
What you can build

A general-purpose application platform.

Plus multi-site hosting, file management, localization, mail templates, spatial indexing and Server-Sent Events.

  • Internal business apps
  • CRM & ticketing
  • Asset tracking
  • Customer portals
  • Data platforms with complex access control
  • Document & content management
  • Multi-source integrations (JDBC, MongoDB, Kafka, MQTT, REST)
  • Workflow automation
  • Prototypes that grow into production
How it works

One system instead of a stack of tools.

Get started with zero headache and zero dependencies: No database, ORM, framework, API, Auth tool needed - Structr has it all out of the box.

  • Install Structr

    Setup

    Start a full instance with Neo4j in one command, or use a free cloud instance.

    docker compose up -d

  • Define your data model

    Types

    Types, properties, relationships and constraints – modeled in the visual schema editor and enforced at database level.

    Report → HAS_STATE → State

  • Get your REST API immediately

    Data

    Every type is a CRUD endpoint with filtering, sorting and pagination. JSON views control what each context sees.

    GET /structr/rest/Report?state=open

  • Build your first page

    Screens

    Server-side rendering with template expressions, repeaters and event–action mappings straight to backend operations.

    ${report.title}

  • Automate your first workflow

    Processes

    Draw BPMN processes in the visual process editor: user tasks, decision gateways, timers and messages – bound to your types and assigned to the right people.

    Submit → Review → Decision? → Publish

  • Activate the built-in security

    Permissions

    Users, groups, object-level permissions, OAuth 2.0, JWT, 2FA and SSL are core – down to individual page elements.

    $.grant(staffGroup, $.this, 'read, write')

  • Write your business logic

    Power Tools

    When the visual tools stop, server-side JavaScript on GraalVM and StructrScript continue. Methods, lifecycle callbacks, cron jobs.

    $.find("Report", { state: "open" })
Schema changes are instant.

The schema is graph data itself. Modifications apply in milliseconds while the system is running – no migration scripts, no deployment cycle.

Screenshots

  • The visual schema editor showing the forum types and their relationships

    Schema Editor. Types, properties and relationships – changes take effect in milliseconds.

  • The BPMN process editor showing a forum post review with user tasks, a decision gateway and end events

    Process Editor.Model BPMN processes with user tasks, gateways and events – and run them on your own data.

  • The Pages area with the page tree and the App Graph showing elements, repeaters and wired actions

    UI Builder.Build pages in the page tree and see the App Graph: elements, repeaters, components and wired actions at a glance.

Why a graph database

Relationships are stored, not reconstructed.

Relational platforms rebuild every relationship at query time through JOINs – and the cost multiplies as your model grows. In Neo4j a relationship is a direct pointer: one hop costs the same whether you hold a thousand records or a hundred million. No other major application platform is built this way.

MATCH (r:Report)-[:IN_STATE]->(s:State {name: 'open'}) RETURN r LIMIT 10 · 10 of 234

  • Faster development speed. Connected models need no denormalization, JOIN tuning or query optimization.

  • No overhead. No ORM mismatch: application objects are graph nodes and relationships, and Cypher is available in your business logic.

  • Enterprise-ready. Permissions follow organizational hierarchies and resolve efficiently at scale.

  • Maps real-world data flows. Supply chains, workflows, knowledge graphs and org charts: graph in, graph out.

How Structr compares

Build the 20% that is yours. The platform carries the other 80%.

  • vs.

    Traditional stacks

    React + Node + Postgres, Django, Spring, Rails: one platform replaces the whole assembly, and its maintenance surface.

  • vs.

    Other application platforms

    Self-hosted, standard JavaScript instead of proprietary languages, no tier limits on users, apps or API calls – and a graph underneath.

  • vs.

    Off-the-shelf software

    Model your own processes instead of bending them to a vendor’s assumptions. Change the schema when requirements change.

  • vs.

    Fully custom development

    Persistence, APIs, auth, users, files and rendering exist on day one. Your team builds the domain-specific part.

  • vs.

    Closed-source platforms

    Audit it, extend it, fix it. For regulated industries this is risk management, not philosophy.

  • vs.

    US-owned platforms

    GDPR and data residency without CLOUD-Act exposure. Developed and maintained under German and EU law.

Data residency. Self-hosted means your data stays exactly where you deploy it. No telemetry, no cloud dependencies.

Legal certainty. No Schrems-II uncertainty around transatlantic transfers – the vendor sits under EU jurisdiction.

Geopolitical independence. Unaffected by sanctions, export restrictions or policy changes abroad.

Long-term alignment. 15+ years as an independent German Mittelstand company – no VC pressure to pivot or exit.

Technical foundation

Standard parts. Nothing proprietary in the way.

Your data sits in a standard Neo4j database you can query, back up and migrate independently of Structr.

  • Java
  • GraalVM
  • Neo4j
  • Apache Pulsar
  • DockerDocker
  • Debian
  • MongoDB
  • RuntimeJava on GraalVM – Java, JavaScript, Python, Ruby and more
  • DatabaseNeo4j (primary); experimental: Memgraph, Amazon Neptune
  • DeploymentDocker, Debian packages, manual install on any Linux server
  • APIsREST, OpenAPI, Server-Sent Events, WebSocket
  • IntegrationsJDBC, MongoDB, MQTT, Kafka, Apache Pulsar, FTP, Email/IMAP, RSS
  • SecurityOAuth 2.0, JWT, JWKS, 2FA, object-level permissions, SSL/TLS
  • ScriptingServer-side JavaScript (GraalVM), StructrScript, Cypher
  • Admin UIDashboard, Schema, Pages, Data, Code, Graph, Flows, Processes
  • RequiresJava 25+, Neo4j 5.x or 2026.x
Get started

Three ways to get started.

  • Try on Cloud

    A free cloud-hosted instance managed by the Structr team. No installation, no credit card, ready in minutes.

    https://structr.cloudTry it now

  • Run on Docker

    The Docker setup gets you a full instance with Neo4j for development.

    docker compose up -dDocker Setup

  • Deploy on Linux

    Structr installs as a standard Debian package on any Linux server. Then build your first application with the tutorial.

    Debian package · any Linux serverBuild Your First Application