Build a hierarchy
Add components at any level. A platform can contain applications, an application can contain services, and a service can contain individual components.
ONE MODEL. EVERY LEVEL.
Map your architecture, step inside any component, and follow the connections.
Define the language that makes sense for your world.
A connected model,
from overview to implementation.
Double-click a component to explore inside.
02 / MAKE IT YOUR OWN
Define component types, colours, shapes and custom properties. Add relationship types with optional rules about what they can connect.
Platform, application, data store…
or something entirely your own.
Depends on, calls, reads / writes…
connect any level to any other.
Owner, lifecycle, classification…
add the details your model needs.
03 / A LITTLE ORIENTATION
Add components at any level. A platform can contain applications, an application can contain services, and a service can contain individual components.
Double-click a block, press Enter on it, or choose “Explore inside” in Details. Use the breadcrumb to return. Expand blocks to compare several levels in the same view.
Choose any source and target in the relationship editor. Links roll up to visible parents when collapsed. Select a link to see every original endpoint.
Select a component to inspect its connections or trace incoming and outgoing links. Traces follow explicit directed relationships; containment does not imply a dependency.
This is an independent, configurable hierarchical modeller. It does not implement the ArchiMate specification, its full notation or exchange format. Define your own conventions and relationship rules, and use the included examples as a starting point.
The view includes external context blocks for relationships crossing the current boundary. A collapsed endpoint is represented by its visible ancestor. Multiple links with the same visible endpoints and relationship type are grouped; selecting a group reveals its exact links. Links internal to a collapsed component remain in the model and reappear when expanded.
The active model and its component positions are saved in this browser’s local storage when available. It is a single-user tool with one active model per browser origin. Export JSON before clearing browser data or changing devices. Browser storage and file:// persistence vary by browser; the saved-state message explains any failures. There is no account, shared backend or real-time collaboration in this edition.
The file format accepts up to 3,000 components, 10,000 relationships, 64 nesting levels and 50 types of each kind. Imports are limited to 5 MB. These are validation limits, not a performance guarantee: use focused and collapsed views for large models. Types can define up to 16 custom property fields. The initial renderer uses a deterministic grid arrangement with draggable blocks; automatic routing and overlapping manually placed links may need adjustment.
04 / YOUR TOOL, ON YOUR TERMS
Download the same modeller as a standalone HTML file. Open it straight from your computer, without installing anything. Import your model JSON to pick up where you left off.
No installation · No model upload · Generated from the same maintained source
05 / NETWORK TRANSPARENCY
Diagram editing, validation, layout, imports and exports run locally. The application has no model-upload endpoint, analytics or telemetry. The online edition loads Mightora branding, shared components and the YAML parser from external domains. Following external links opens those sites.
The offline edition embeds its code, styles and logo, and uses a restrictive Content Security Policy to block application network connections. Downloading a file online or following an external link is an explicit network action.
On a computer, press F12 (or use your browser’s Inspect menu), open Developer Tools → Network, enable Preserve log and reload. Make a model change and inspect the requests. Some computers require Fn + F12; on macOS you can use ⌥⌘I. Mobile browsers may require remote inspection.
The resource summary below uses the Performance Resource Timing API. It may omit blocked requests, some cross-origin details, extensions, cached resources or activity outside this page. It is not a substitute for Developer Tools or an independent audit.
Reading browser resource timing…
Resource timing only. No model content is collected.
MODEL ONCE / SHARE MANY WAYS
Use Export YAML for readable, editable model text, or Export → Model JSON. Import accepts .json, .yaml, .yml and YAML-formatted .txt files. Both formats preserve IDs, custom types, properties, product icons and named views.
Choose a focus, expand the blocks you need, then use Save view. Relationship filters and external context settings are stored with the layout and zoom. Manage views lets you rename, delete and reorder them.
Choose Export Office pack, select your views and generate a .pptx. Each view becomes a slide with editable blocks, labels and attached connectors. Product icons remain embedded images. Exact relationship endpoints appear in speaker notes.
Choose Word in the export panel. Each view starts a new section with a scalable SVG diagram and a PNG fallback. Descriptions and optional component and relationship tables are editable; diagram elements are embedded images.
The icon picker includes Power Apps, Power Automate, Dataverse, Power Pages and Power BI from the official archived Microsoft PowerBI Icons repository. These are the versions supplied by that archive and may differ from current branding. Microsoft and contributors license repository content under CC BY 4.0. Raster fallbacks are generated at the original proportions; SVG artwork is retained.
Use product icons to accurately identify those products in architecture diagrams. Keep product names nearby, preserve proportions, and do not crop, rotate, distort or present them as Mightora’s own branding. The application’s MIT licence does not replace the icon terms or grant trademark rights. See Microsoft’s architecture icon guidance and the source repository’s legal notices.
ONE MODEL / MULTIPLE DIMENSIONS
Switch between solution architecture, business processes, stakeholder responsibilities, application layers, business traceability and all entities. Use Scope to filter a process and its nested steps or a component layer. Save each perspective, then select several views for a Word or PowerPoint pack.
Add a Process, then add Child steps. Add Stakeholders and use Business mappings to record their participation and RACI responsibilities. Connect process steps with sequence links and map supporting components to the exact process or step. This is lightweight process mapping, not BPMN execution.
Use Application layers to define names, responsibilities, colours and order. Edit components to assign their primary and optional additional layers. Layers are independent of nesting; the layer view shows each component once in its primary layer, or its first matching layer when filtered.
Business traceability brings processes, stakeholders and supporting components together. RACI roles are descriptive rather than policy-enforced; assignments to whole processes are not silently inherited by steps. Try the business-process example to see all dimensions already mapped.
YOUR MODEL / YOUR WORKSPACE
Use the same modeller in desktop VS Code, vscode.dev and browser-based Codespaces. Save JSON or YAML alongside your source, review model changes in Git, and export diagrams through VS Code file pickers. The extension bundles the application and makes no analytics requests.
Version 0.4.0. Marketplace publication is a separate step. Interactive VS Code web and desktop verification is still required before a Marketplace release.
This tool is open source under MIT and can be independently inspected. Workflow links show actual results; no security certification is implied.