AI Integration

Schematify graphs are structured data: nodes, relationships, attributes, and hierarchy. That makes them a good fit for AI agents. An agent can inspect a codebase, read a design, or follow a written description, then produce a graph that can be pushed to Schematify.

AI integration has two parts: the Schematify CLI and Agent Skills.

How it works

The CLI manages graphs on your Schematify server. It can list, pull, push, delete, validate, convert, and publish graph data.

Agent Skills teach your AI assistant how to use the CLI and how to build useful Schematify graphs. With the skills installed, you can ask your agent to create or update graphs as part of a normal coding session.

Getting set up

1. Run the installer

Copy the one-liner from the Document Manager’s CLI & Agent Setup page or the CLI installation guide. The installer installs the CLI, then prompts to install AI agent skills (Yes recommended). No Node.js is required.

For the hosted Schematify service, setup is not required. Authenticate with:

schematify login

Run setup only if you are using a custom VM or another non-default deployment:

schematify setup
schematify login

2. Ask your agent for a graph

Examples:

• “Create a Schematify architecture graph of this project”
• “Generate a dependency graph for apps/client/src”
• “Map this deployment topology as a Schematify graph”
• “Pull the existing infrastructure graph, add the new service, and push it back”

Available skills

Skill	Purpose
schematify-cli	Teaches the agent to operate the CLI, inspect command help, use JSON output, and manage graphs on the server.
schematify-graphs	Teaches the agent to author and edit Scriptable Graph TypeScript using the graph builder API, channels, and live channel publishing.
schematify-generate	Teaches the agent to generate architecture graphs, dependency graphs, or custom graphs from a project, directory, dataset, or description.
schematify-render	Teaches the agent to configure specialized node rendering such as property cards, reports, and charts.

The skills compose: generation decides what should be represented, graph authoring turns that into a Scriptable Graph, render guidance configures specialized displays, and the CLI runs or publishes it.

Typical workflow

1. You ask for a graph or an update to an existing graph.
2. The agent decides whether this is an architecture graph, dependency graph, or custom graph.
3. The agent scans the relevant files or follows your description.
4. It authors or updates a Scriptable Graph.
5. It runs the graph with the CLI to validate or publish it.
6. You open the graph in the Graph Viewer.

For existing graphs, the agent can use the CLI to pull the current document, make a targeted change, and push it back.

What’s next

• Agent Skills covers the skills and usage examples.
• CLI is the command reference for the Schematify CLI.
• Scriptable Graphs covers TypeScript-based graph generation.