Architects spend months building rigorous models in Sparx EA. That work disappears the moment someone closes the EA client. Stakeholders, executives, and even other architects don’t see it. They see whatever was in last month’s PowerPoint.
EA GraphLink changes that. It’s a connectivity layer that transforms your EA Repository into a live data source for every AI tool your organization uses. One deployment connects Power BI, Copilot, Fabric, Tableau, Agentforce, Claude, Cursor, and more: simultaneously, on the same data.
The catch is important: EA GraphLink doesn’t invent structure where none exists. It transforms your EA Repository using the MDG Technology definition: the formal semantic schema you’ve already built (or need to build). A well-governed MDG produces rich, queryable architecture intelligence. A fragmented one produces confused outputs at scale. MDG quality is assessed in Discover and established in Deploy. It’s the upstream prerequisite for everything downstream.
When you deploy Sparx EA, you solve the first problem: architects get a single source of truth for models. But you create a second problem: that truth is locked inside the EA client, visible only to people with EA licenses who know how to navigate the interface.
Stakeholders have questions that architecture data could answer:
But those questions go to architects via email or Slack. Architects export reports manually. Decision-making happens against stale snapshots. Data governance teams build their own parallel copies to feed BI tools. Integration teams replicate capabilities in their platforms because they can’t access the authoritative source.
When EA Repository is live-connected:
That’s what EA GraphLink enables.
EA GraphLink sits at the center of the Sparx Systems Architecture Platform stack. It connects directly to the EA Repository database (bypassing the EA client and Pro Cloud Server), transforms the physical schema using the MDG Technology definition, and exposes two interfaces: a GraphQL API for BI dashboards, and an MCP Server for AI tools and data integration platforms.
The Architecture:
| Component | Role |
|---|---|
| EA Repository | Physical database (SQL Server, MySQL, PostgreSQL) containing all model data |
| EA GraphLink | Connectivity and transformation layer: reads repository, applies MDG schema, exposes interfaces |
| GraphQL API | Interface A: Returns architecture data in structured format for Power BI and Tableau connectors |
| MCP Server | Interface B: Exposes architecture as MCP resources for Claude, Copilot, Agentforce, Cursor, and data integration control planes |
One EA GraphLink deployment supports both interfaces simultaneously. You deploy once, and all downstream tools can connect.
EA GraphLink can only expose what the MDG Technology definition describes. The MDG is the semantic bridge between the physical repository schema and the logical architecture model.
This is why Discover (assessment) and Deploy (governance establishment) come before Connect. You cannot skip the MDG foundation.
Different organizations have different downstream tools. EA GraphLink adapts.
If your organization standardizes on Microsoft 365 and Azure:
If your organization standardizes on Salesforce and MuleSoft:
Regardless of ecosystem:
| Integration | Type | Who Benefits | Ecosystem | Start Here If… |
|---|---|---|---|---|
| Power BI | BI Dashboard (GraphQL) | Executive, Analyst, Architecture Manager | Microsoft | You use Power BI for executive dashboards and want live architecture visibility |
| Microsoft Copilot | AI Chat (MCP) | Architect, EA Architect, Stakeholder | Microsoft | You want stakeholders to ask architecture questions in Copilot |
| Microsoft Fabric | Data Integration (MCP) | Data Engineer, Data Architect | Microsoft | You’re building a enterprise data fabric and architecture is a first-class source |
| Tableau | BI Dashboard (GraphQL) | Executive, Analyst, Architecture Manager | Salesforce | You standardize on Tableau for BI and want live architecture visualization |
| Salesforce Agentforce | AI Agent (MCP) | Sales, Service, Executive, Stakeholder | Salesforce | You want Agentforce agents to reference architecture in customer and operational conversations |
| MuleSoft Fabric | Data Integration (MCP) | Integration Architect, Data Engineer | Salesforce | Architecture data is a required source in your MuleSoft integration platform |
| Claude | AI Chat (MCP) | Architect, EA Architect, Software Engineer | Universal | You use Claude for architecture analysis and want live repository context |
| Cursor / Claude Code | IDE Integration (MCP) | Software Architect, Developer | Universal | You want IDE-native access to live architecture while coding or designing systems |
| Google Gemini | AI Chat (MCP) | Architect, EA Architect | Universal | Your team uses Gemini and wants architecture context in chat |
| ChatGPT Enterprise | AI Chat (MCP) | Architect, EA Architect | Universal | You standardize on OpenAI and want architecture context in enterprise ChatGPT |
| Azure OpenAI | AI Chat (MCP) | Architect, EA Architect, Enterprise Governance | Universal | You deploy OpenAI models in Azure and need architecture context without external API calls |
| Kernaro AI Hub | Stakeholder Platform | Business Stakeholder, Executive, EA Architect | Universal (Sparx-hosted) | You want non-technical stakeholders to ask natural language architecture questions in a browser |
Start with: EA GraphLink Foundation → Power BI → Copilot
Timeline: 8-12 weeks. Cost range: $50K–$185K+ (Connect service).
Start with: EA GraphLink Foundation → Tableau → Agentforce
Timeline: 8-12 weeks. Cost range: $50K–$185K+ (Connect service).
Start with: EA GraphLink Foundation → Integration Comparison
Timeline: 10-16 weeks. Cost range: $75K–$185K+ (Connect service with extended testing scope).
CRITICAL: Do not start a Connect engagement if your MDG Technology definition is in poor shape.
Your MDG is in good shape if:
Your MDG needs work if:
If your MDG is strong: Proceed directly to Connect.
If your MDG needs work: Start with Deploy ($30K–$130K). Deploy establishes MDG governance, documents definitions, and ensures consistent application across the repository. Then Connect.
If you’re not sure: Start with Discover ($25K–$75K). Discover includes an MDG readiness assessment and a roadmap for governance work.
EA GraphLink is a connectivity and transformation layer that connects your EA Repository directly to Power BI, Tableau, AI tools, and data integration platforms. It reads data from the physical repository, transforms it using your MDG Technology definition, and exposes two interfaces: a GraphQL API for BI tools and an MCP Server for AI tools.
No. One EA GraphLink deployment serves both. The GraphQL interface connects to Power BI and Tableau simultaneously. The MCP Server connects to Copilot, Agentforce, Claude, and every other tool at the same time. Deploy once, integrate many.
The MDG Technology definition is the semantic schema that describes your architecture data. EA GraphLink transforms the physical repository schema into a logical model using the MDG. If the MDG is fragmented or informal, EA GraphLink cannot produce meaningful outputs: not because of a software limitation, but because there is no semantic foundation to transform. MDG quality is assessed in Discover and established in Deploy before you Connect.
Yes. Claude, Cursor, Claude Code, Google Gemini, ChatGPT Enterprise, and Azure OpenAI all connect via the MCP Server interface. EA GraphLink uses the open MCP standard: you’re not locked to one vendor’s AI platform.
GraphQL serves BI tools (Power BI, Tableau) that need structured data for dashboards. MCP serves AI tools and data integration platforms that need to understand and reason about architecture. Both interfaces expose the same underlying data, transformed by the same MDG. You use one or both depending on your tools.
8–16 weeks, depending on scope. A simple deployment (EA GraphLink + one BI tool) is faster. Multi-tool integrations with testing and optimization take longer. Sparx Services works with you on timeline during scoping.
You purchase EA GraphLink, Kernaro Hub, and any Sparx Systems software directly from Sparx Systems. Sparx Services provides the software bill of materials as part of the Connect engagement. You purchase Power BI, Tableau, Copilot, and Agentforce licenses directly from Microsoft or Salesforce. Sparx Services is not a reseller: we help you understand what you need and how to use it.
Start with Deploy ($30K–$130K) to establish MDG governance and document definitions across your repository. Once governance is in place and consistent application is verified, then move to Connect. A Connect engagement on top of a fragmented MDG will produce unusable results; a Connect engagement after Deploy produces high-value integrations.
Your architecture models contain insights that should drive decisions across the business. EA GraphLink makes those insights live, queryable, and accessible to stakeholders, executives, and AI tools: without exporting, copying, or manual reporting.
Three paths forward:
[CTA Button: Schedule a Consultation] | [Learn about Discover] | [Learn about Deploy]
Talk to a Sparx Services architect about where your organization is on the journey and what the next stage looks like.