Nuclos MCP
AI in ERP development, with control and traceability. For development teams, IT departments, and partners.
MCP (Model Context Protocol) connects AI agents and your development environment to the data model, rules, and project artifacts. Model, version, and accelerate with AI, inside the proven low-code frame.
- MCP in plain language
- Sparring · validation · gates
- Open source & low-code
What is MCP?
Structured AI access for ERP development.
MCP (Model Context Protocol) (opens in new tab) is an open protocol that lets AI assistants and development tools such as Cursor access defined tools and data sources. The agent reads metadata, rules, and configuration directly from the Nuclos development environment, not from screenshots or PDF exports.
Typical use in Nuclos 5: model data and rules, design UIs, prepare reviews, and generate deployment artifacts. MCP accelerates development and configuration on the platform.
MCP does not replace your pipeline. It supplies context and proposals. Review, validation, and quality gates stay with your team and your Git/CI process chain.
- Sparring AI proposes, you correct and approve.
- Validation Rules and platform patterns before every change.
- Quality gate Review and approval before test.
IDE · MCP
Development in dialogue
Waiting for dialogue …
Sparring proposal
First draft
- Field
- Required
- Scope
- Approval
Validation before adoption
- Agent rules followed Proposal stays within fixed operating rules.
- Low-code patterns validated No structure outside the metadata model.
- No direct database access Change as configuration artifact, not a schema jump.
- Metadata within project scope Only entities and rules of the agreed sub-process.
Approval before test
Change: mandatory validation · order entity
- Change versioned Configuration is stored as a traceable project artifact.
- Review by IT Your team checks scope, rules, and impact.
- Rules verified Project policies and quality gates are met.
- Approved for test Release role authorizes promotion to the test environment.
IT reviews · release approves
Simplified demo. Sparring, validation, and gates are defined bindingly in your project.
Why is a chatbot not enough?
Speed without control is not enough.
Many teams try AI for code and configuration. Without guardrails, proposals drift outside the data model, leave no review trail, and skip gates before production. IT and business lose a shared basis for decisions.
Nuclos 5 combines AI acceleration with low-code boundaries (opens in new tab), fixed agent rules, and the familiar chain of Git, review, and deployment. The result stays maintainable, auditable, and under your control.
Where AI proposes, you decide
Sparring, validation, and quality gates as one system.
-
Sparring partner in modeling
The AI agent proposes changes to the data model, rules, and UIs. You discuss, correct, and approve only when business and IT perspectives align. AI does not replace architecture decisions; it speeds up preparation.
-
Validation before every change
The agent follows fixed operating rules and stays within validated platform patterns. No direct access to the database structure. Proposals remain inside the low-code frame, traceable for IT and compliance.
-
Quality gates before deployment
Every change is stored as a versioned project artifact in Git. Pull request review, CI hooks, and gate-controlled promotion from dev through test to production remain mandatory before anything goes live.
Benefits at a glance
Why teams choose Nuclos MCP over isolated tool chains for individual ERP development.
-
One platform instead of a tool chain
Low-code, runtime, integrations, AI, and deployment are unified. No media breaks between model, code, and operations.
-
Speed with control
AI accelerates modeling and documentation. Git and review processes ensure quality and traceability.
-
Partner and enterprise ready
Multi-tenancy, roles, and audit for consultancies and regulated environments.
-
Future-proof with MCP
Model Context Protocol connects AI agents and IDEs to modeling and metadata. It enables AI-assisted ERP development, not chat over live ERP data.
-
No license trap
Open-source foundation: customization and operations stay in your hands, with transparent packages instead of hidden module costs.
-
Clear path to Nuclos Enterprise
From prototype to production delivery in 30 days with Nuclos experts when you want to outsource execution speed.
Integrations & automation
Nuclos MCP connects development, project control, and operations in one environment.
Git integration
Every change to the model and extensions is versioned, like professional software development.
- Connect GitHub, GitLab, Bitbucket (Enterprise: self-hosted)
- Branches per feature, release, or client project
- Pull-request workflow with review before merge
- Tags and releases for production states
- CI hooks for automated builds and tests
Jira integration
Requirements and development stay linked from epic to deployed build.
- Link Jira issues to Nuclos artifacts
- Status sync between Jira issues and development progress
- Traceability for audits and project controlling
- Sprint and release planning with deployment targets
- Optional: Confluence links for specifications
Deployment automation
From development to target environment, repeatable, documented, and rollback-capable.
- Pipelines for dev, test, and production
- One-click or gate-controlled promotions
- On-premise, private cloud, or Nuclos hosting (DE)
- Migration and configuration packages
- Automated rollback and release notes
AI capabilities
AI as a development accelerator, not a replacement for your architecture decisions.
- Assistants for modeling, rules, and UIs
- Code and documentation suggestions with project context
- AI-assisted test data and test scenarios
- AI-assisted conversion of workshop requirements into structured artifacts
- Enterprise: custom models / data residency by agreement
MCP in the dev context
Model Context Protocol connects AI agents and development tools to platform configuration in a controlled and auditable way.
- MCP interfaces for data model, rules, UIs, and project metadata
- Agents and IDEs (e.g. Cursor) work in the development context, not on live ERP operations
- Standardized tools for modeling, reviews, and deployment preparation
- Partners: custom MCP extensions for client development projects
- Compatible with the MCP ecosystem for developer workflows
Pricing
Nuclos MCP for development, with clear packages instead of hidden module costs.
Placeholder pricing, final terms in the quote. All figures are non-binding, excl. VAT. Individual quotes for partners and enterprises.
Developer
from €XXX / month
IT teams, developers, and partners
MCP access with low-code, Git, and AI assistance for self-build.
- MCP access for AI agents and IDEs (e.g. Cursor)
- Low-code modeling with Git integration
- AI assistants for modeling and documentation
- Standard support and onboarding
Partner
Custom
Partners, enterprises, regulated environments
Self-hosted, custom MCP, and compliance by agreement.
- Self-hosted Git, custom MCP included
- On-premise, private cloud, SSO & compliance
- Dedicated success & architect
- Custom AI models / data residency optional
Put MCP and AI to work in your ERP development
Discuss architecture, MCP connectivity, and quality gates with our team, or take a look at the packages.
Sparring · validation · gates · MCP in the dev stack · Open source & low-code