TrainAR Documentation
Everything you need to get started with TrainAR - from platform basics to API integration and MCP server configuration.
Getting Started
What is TrainAR?
TrainAR is an augmented reality training platform built for the construction industry. It combines AR smart glasses, a cloud-based management platform, and AI-powered training tools to help construction businesses train engineers faster, reduce errors, and maintain compliance automatically.
The platform includes skills management, a knowledge base, training session recording and review, performance reporting, integrations with field service tools, and a full REST API with MCP server access for AI agents.
How It Works
Create Skills
Build API and RAG skills that enhance training sessions - fetch parts diagrams, boiler manuals, and reference material on demand.
Train with AR
Engineers wear smart glasses that overlay guidance onto real equipment. Voice commands keep hands free for actual work.
Track and Report
Every session is recorded with video, voice and step data. Dashboards show session activity, completion rates, time-on-task and minute usage by user, team or trade skill.
Platform Guide
Skills
Build and manage API and RAG skills that plug into live training sessions. Skills let the AI fetch exploded parts diagrams, manuals, and compliance docs on demand.
Knowledge
Build your organisation's knowledge base. Upload documents, procedures, manufacturer guides, and compliance requirements. The AI indexes everything for instant retrieval.
Sessions
Training sessions capture video, voice, procedures used and the session summary. Trainer + trainee role split, with status for started / completed / failed. Drill into any session for the recording and details.
Reports
Sessions, minutes, tasks and users reports. AR session activity, completion rates, time-on-task and AI minute usage broken down by user, team or trade skill.
Compliance
Automatic documentation of every training activity with timestamped evidence. Generate audit-ready compliance reports for regulators and clients.
Settings
Configure your organisation - branding, user roles, integrations, notification preferences, and data retention policies. White-label options available on Professional plans.
API Reference
REST API
The TrainAR REST API provides full programmatic access to the platform. All endpoints return JSON and use standard HTTP methods.
Base URL: https://xbzrvdacvnodvpplkgyv.supabase.co/functions/v1 Authentication: Bearer token (API key prefix: tak_) Header: Authorization: Bearer <your-api-key>
Available Endpoints
/api-skills/api-skills-execute/api-knowledge/api-knowledge-process/api-tenant-sessions/api-tenant-tasks/api-tenant-tasks/api-tenant-tasks/api-tenant-users/api-reportsAuthentication
API keys are generated from your organisation settings in the platform. Each key is scoped to your organisation and can be rotated at any time.
curl -X GET \ https://xbzrvdacvnodvpplkgyv.supabase.co/functions/v1/api-skills \ -H "Authorization: Bearer tak_..." \ -H "Content-Type: application/json"
Rate Limits
MCP Server
What is the MCP Server?
The TrainAR MCP (Model Context Protocol) server exposes 70 tools that allow AI agents to interact with the full TrainAR platform. Agents can create skills, manage knowledge, run reports, handle billing, connect integrations, and more - all through a standardised protocol.
MCP is an open standard for connecting AI models to external tools and data sources. TrainAR's MCP server works with Claude Desktop, Claude Code, and any MCP-compatible client.
70 Tools Available
The MCP server provides tools across every area of the platform:
Connect to Claude Desktop
Add this to your Claude Desktop configuration file to connect to the TrainAR MCP server:
{
"mcpServers": {
"trainar": {
"command": "npx",
"args": [
"-y",
"@anthropic-ai/mcp-remote",
"https://mcp.trainar.ai/sse"
]
}
}
}Agent Signup
AI agents can sign up for TrainAR access using the agent_signup tool. No human intervention required - the agent receives API credentials and can start managing the platform immediately.
// Agent calls the agent_signup tool with:
{
"organisation_name": "Acme Plumbing Ltd",
"admin_email": "admin@acmeplumbing.co.uk",
"plan": "professional"
}