Frequently asked questions
Short answers. The documentation folder in the repository has the long ones.
What is ApexDiagram?
A network diagramming app where the diagram is a database record, not a drawing. Every device, port, cable, VLAN and rack slot is stored once and rendered as five views: Physical, Network, Logical, Data flow and Purdue. Edit a device in any view and the others update.
How is that different from Visio, draw.io or Lucidchart?
Those are drawing tools: shapes and lines with no meaning. ApexDiagram knows that a line is a Cat6 cable from port 3 of one switch to port 12 of another, so it can lint the network, trace a cable through a patch panel, roll up a bill of materials, plan subnets and export a lab topology. You can still import your existing draw.io and Visio files.
Who is it for?
Small and medium businesses that keep the network in a binder, MSPs that look after many of them, OT and industrial teams that need Purdue and IEC 62443 views, and home-lab owners and students on the free plan.
What can I import?
draw.io (.drawio/.xml), Visio (.vsdx), CSV and Excel device lists, LLDP/CDP neighbour tables, pfSense and OPNsense config.xml, nmap XML, UniFi and Meraki JSON exports, NetBox JSON or a live NetBox pull, and ApexDiagram's own JSON and YAML. Drop the file on the dashboard; the format is detected from the content.
Does ApexDiagram scan my network?
No. There is no agent or poller. Discovery data comes in as files or through the API, and the Refresh from file page shows a diff before anything is written. Scheduled pulls can call the same API.
What is the AI architect?
An assistant that edits the diagram from a description: “add a second core switch and cable every IDF to both”. It works through the same service layer as the REST API, every run saves a version first, and every change it makes is in the audit trail. It needs an Anthropic API key configured on the server. See the privacy notes for what is sent.
Is there an API?
Yes. A REST API under /api/v1 with an OpenAPI 3.1 description, and an MCP server at /api/mcp so AI assistants can read and edit diagrams. Keys are created at Settings → API keys and carry a role.
How do roles and approval work?
Viewers read, editors edit, admins manage members, keys and approvals, owners manage the organisation. A diagram can require approval: editors then propose a change request, which is a full draft copy, and an admin reviews the diff and approves it into the live diagram. Everything is logged in the Activity tab.
Can I share a diagram with someone who has no account?
Create a share link from the diagram menu. It is read-only, can expire, and can be revoked. Presentation mode steps through the views full screen.
Where is my data stored and can I get it out?
In a PostgreSQL database for your organisation. Export any diagram as JSON or YAML at any time; that file re-imports losslessly. Images export as PNG, SVG and PDF; BOMs as CSV and Excel; topologies as Containerlab YAML.
Is there a free plan?
Yes: one person, five diagrams, every view, importer and the linter, plus a monthly allowance of AI architect runs. Collaboration, change approval, audit export and unlimited diagrams need Team. See Pricing.
Does it work on a phone?
Reading and inspecting, yes. Palettes and side panels collapse on small screens and inspectors open as a bottom sheet. Drawing is best on a laptop.
What leaves the server?
Nothing, unless you use a feature that calls out: the AI architect sends the current diagram summary and your prompt to Anthropic; a NetBox pull contacts the NetBox URL you give it with the token you give it, and the token is not stored. Details are in the privacy policy.