Dev Portal

Nuclos for Developers and IT Teams

Open architecture, REST API, MCP integration, and AGPL licence.

This page is for IT leaders and developers evaluating Nuclos on technical merit. You will find the platform architecture, the REST API, MCP integration, and the way to the wiki and repository.

  • Open Source · AGPL 3.0
  • REST API from data model
  • MCP integration
  • On-premise or cloud

Architecture at a Glance

Five layers from the metadata model to deployment: open, traceable, no black-box system.

  1. Metadata model

    Every object, field, relationship, and rule is defined as metadata. The data model is fully configurable, versioned, and forms the base for every layer above it.

  2. Low-code layer

    Screens, workflows, and business logic are built in the low-code editor on top of the metadata model. AI agents have no direct database access: the solution space stays within validated platform patterns.

  3. REST API

    Every modelled object automatically generates a consistent REST API: GET, POST, PUT, DELETE. One schema, directly testable, no manual endpoint maintenance.

  4. MCP integration

    MCP (Model Context Protocol) connects AI agents directly to the ERP system's processes and data. Agents can read, write, and trigger actions, within the boundaries of the low-code layer.

  5. Deployment

    Nuclos runs on-premise on your own infrastructure or in the Novabit Cloud (made and hosted in Germany). The source code stays under AGPL 3.0. No vendor dependency.

REST API and Integration

Generated automatically from the data model: one schema, full CRUD.

  • Endpoints are generated automatically from the data model, no manual upkeep.
  • Consistent schema: GET, POST, PUT, DELETE for every modelled object.
  • Authentication via API token or session. Details in the wiki.
  • Full test coverage via Swagger/OpenAPI. Reference in the wiki.
GET /api/v1/auftrag/{id}
Authorization: Bearer <token>

→ 200 OK
{
  "id": "4711",
  "status": "offen",
  "kunde": { "id": "200", "name": "Muster GmbH" },
  "positionen": [ … ]
}

MCP and AI Development

AI agents work within your ERP processes, inside the low-code layer's boundaries.

MCP (Model Context Protocol) (opens in new tab) connects AI agents directly to Nuclos and lets them act within your ERP processes: reading data, triggering workflows, writing back results.

The low-code layer defines the boundary. Agents have no direct access to the underlying data structures. This constraint is intentional: it keeps the solution space within validated platform patterns and makes AI-generated results maintainable and predictable.

  1. Provision an instance

    Install Nuclos 5 or provision it in the cloud. On-premise or Novabit Cloud.

  2. Configure the MCP endpoint

    Add the MCP endpoint to your AI agent or IDE, for example Claude or Cursor.

  3. Authenticate with a token

    Access via an API token with a defined role and permissions.

  4. The agent works within its role

    The agent reads, writes, and triggers actions, within the permissions of the token role.

Ready to build with MCP and low-code in Nuclos?

Nuclos MCP is the environment for modelling, configuration, and self-build with AI and MCP.

Open Source · AGPL 3.0

AGPL 3.0 means the source code, its evolution, and its operation stay in your hands.

The source code is yours

The complete Nuclos source code is licensed under GNU AGPL 3.0. No per-user licence fees, no premium tier for core functionality.

No vendor lock-in

If you decide to run or extend Nuclos yourself, the entire codebase is available to you. Any qualified development team can maintain the system.

Source code access

The full source code is available in the repository for transparency and traceability.

View repository

Security disclosure

Report vulnerabilities directly to our team, not publicly.

Get in touch

Deployment and Operations

On-premise or Novabit Cloud: made and hosted in Germany, run under GDPR.

On-premise

Nuclos runs on your own infrastructure. System requirements, supported operating systems, and database requirements are in the wiki.

System requirements in the wiki

Novabit Cloud

A managed cloud environment, made and hosted in Germany. Data stays in Germany, no infrastructure to maintain.

Hybrid

ERP data on-premise, development in the Cloud Workspace. Both run on the same codebase.

Updates

Platform updates follow a defined release cycle. Customisations on the low-code layer stay update-safe, no forked core.

Data residency

On-premise: your infrastructure, your data. Novabit Cloud: servers in Germany, GDPR-compliant.

Support

Technical support through the support portal, with defined response times.

Open support portal

How do you want to work with Nuclos?

Self-build, partner integration, or technical consultation: choose the right entry point.

Open Source · AGPL 3.0 · Made & Hosted in Germany · 20+ years of Novabit experience