Your eConnect Data, Your AI: Custom Reports and Analysis You've Never Had Before
For years, the analysis you could do with your eConnect data was limited to what the application itself offered. The reports were good — but they were our reports, not yours. Built for the general case, not for the exact question your shift supervisor is asking at 2 AM.
That changes today.
Introducing the eConnect MCP Server
We've launched an MCP (Model Context Protocol) server at mcp.econnect.ai that connects AI assistants — ChatGPT Desktop, Claude Desktop, VS Code Copilot, Cursor, and practically any other agentic AI tool — directly to your eConnect system.
What this means in practice: you open your AI client, paste in a token, and start asking questions as if you were talking to someone who has complete access to your databases.
"Find face detections from the last 4 hours and group them by camera"
"Show me POS refund exceptions over $100 from tonight's shift"
"List all LPR hotlist hits from the parking garage this week"
"How many table sessions ran longer than 6 hours in the casino today?"
No SQL. No API calls. No reports to schedule. Just questions and answers, in real time.
What You Can Actually Do With This
Here's what makes the MCP server genuinely different from anything we've offered before.
Build Reports That Never Existed
The eConnect app ships with standard reports because they need to work for everyone. Your business isn't everyone. With the MCP server and an AI client, you define the report.
Need a daily face match summary broken down by watch list, camera zone, and hour of day — formatted as a Markdown table and saved directly into a file on your desktop? Just ask. Using VS Code with Copilot in Agent mode, your AI can pull the data and write the document in one shot:
"Get all face watch list matches from today, group by watch list name
and hour, and write a formatted summary into today-face-summary.md"
Cross-Module Analysis That Was Never Possible Before
The MCP server gives a single AI session access to all your enabled modules simultaneously. That opens up cross-module analysis that has never been practical before.
Face + POS correlation:
"Were there any POS void exceptions within 15 minutes of a watch list
face match at the same register area today?"
LPR + Casino:
"Show me all vehicles that entered the parking structure during the
same time window as any flagged casino cage transactions today"
Shift intelligence:
"Give me a full shift summary — face detections, LPR hotlist hits,
POS exceptions, and active table sessions — for the overnight period
midnight to 8 AM"
This kind of cross-referencing previously required either custom development work or a data warehouse. Now it's a single sentence.
Module Breakdown
🧠 Face Recognition
Query enrolled subjects, drill into detection history, filter by camera, time range, watch list, or confidence score. Search by name, ID, or attributes. All 62 MCP tools are available — and the AI knows how to chain them together to answer complex questions.
"How many unique subjects triggered the VIP watch list this week?"
"Find all detections of subject John Smith from last night"
"Which cameras had the most unknown face detections this month?"
🚗 License Plate Recognition
Read plate events, check hot list hits, monitor detector health, and pull plate history for specific vehicles or time windows.
"Has plate ABC 1234 appeared in the last 7 days?"
"Show me all stolen vehicle hot list hits from this week"
"Which LPR detectors are currently offline?"
🛒 POS Analytics
Transaction-level access with exception monitoring, cashier-level drilldowns, register status, and line-item detail. Perfect for building custom exception reports your loss prevention team actually wants.
"Show me every no-sale event from tonight grouped by cashier"
"Get all POS exceptions over $50 from the past week by register"
"Which cashier had the most void transactions this shift?"
🎰 Casino Operations
Table and slot sessions, player profiles, cage transactions, Title 31 compliance events, and aggregated statistics — all queryable through natural language.
"List all table sessions that exceeded $50,000 in action today"
"Show me Title 31 compliance events from the last 30 days"
"Which slot machines had the most sessions this week?"
How Authentication Works
Security was the first thing we focused on when building this. The MCP server logs in as you — using an encrypted token that you generate from mcp.econnect.ai/auth/.
- Tokens are encrypted with AES-256-GCM
- They're valid for 7 days by default (configurable up to 30 days) and are reusable — paste once and it works across sessions until it expires
- The server respects your exact eConnect account permissions — if you can't see a data set in the app, the AI can't see it either
- There's nothing stored server-side about you between sessions
You paste the token into your AI chat at the start of a session. That's it.
Works With the Tools You Already Use
We built the MCP server on the standard Model Context Protocol so it works with the AI tools people are already using:
| AI Client | Notes |
|---|---|
| ChatGPT Desktop | Add via Settings → MCP Servers |
| Claude Desktop | Add via claude_desktop_config.json |
| VS Code / GitHub Copilot | Add via .vscode/mcp.json, use in Agent mode |
| Cursor | Add via Settings → MCP |
| Windsurf | Add via mcp_config.json |
| Custom pipelines | Any MCP client over Streamable HTTP works |
Full setup instructions for each client are in the MCP Server User Guide.
Cloud or On-Premises
The hosted server at mcp.econnect.ai is ready to use right now for cloud-hosted eConnect systems. If you're running eConnect on a closed private network, we can deploy the MCP server container locally — either on the existing eConnect Linux server or a Docker host on your network. Contact support to arrange it.
This Is Just the Beginning
The MCP server launches with 62 tools across 7 categories, and we'll be expanding coverage with every eConnect release. We're also watching closely to see what kinds of analysis our users are actually doing — because those real-world questions will drive what we build next.
If you create something interesting, we'd love to hear about it.
Get started with the eConnect MCP Server →
The eConnect MCP Server is available now at mcp.econnect.ai. Questions or feedback? Reach us at support@econnect.tv or call 702-523-8786.