What if you could manage your entire CRM just by having a conversation with an AI? No clicking through menus. No searching for a customer. No manually updating fields. Just type — or speak — and it happens.
That’s exactly what we’re building with the Webtrix24 MCP Server — and this post covers what it is, what it will do for insurance agents and rental businesses, and how we’re building it on our React, CodeIgniter 3, and MySQL stack.
💡 What is MCP? MCP stands for Model Context Protocol — an open standard developed by Anthropic (the makers of Claude AI) that lets any AI assistant connect directly to software tools and take real actions inside them. Think of it as a universal plug between AI and your business software.
What Is the Webtrix24 MCP Server?
The Webtrix24 MCP Server is a connection layer between your Webtrix24 CRM and any AI assistant that supports the Model Context Protocol — including Claude by Anthropic, Cursor, and other AI tools.
Once connected, your AI assistant can read data from and write data to your Webtrix24 CRM — using plain, conversational language. No technical knowledge required.
For insurance agents — imagine saying:
- “Show me all policies expiring in the next 30 days”
- “Add Rahul Sharma as a new lead with phone 9876543210, interested in motor insurance”
- “How many renewals did we close this month?”
- “Send a WhatsApp follow-up to all leads who haven’t responded in 7 days”
- “What is Priya Patel’s current policy status?”
For computer rental businesses — imagine saying:
- “Which laptops are due for return this week?”
- “Create a delivery challan for TechCorp’s order of 5 Dell laptops”
- “Show me all overdue rentals with customer contact details”
- “Which assets have been idle for more than 15 days?”
- “Generate an invoice for ABC Company’s October rental”
All of this happens instantly, inside your AI assistant, connected live to your actual Webtrix24 data.
Why This Is a Big Deal for CRM Users
CRM adoption fails for one reason more than any other: it takes too many clicks and too much time to do simple things. An insurance agent shouldn’t need to navigate 4 screens to check a customer’s renewal date. A rental manager shouldn’t need to run 3 reports to see what’s available today.
MCP changes this completely. Instead of navigating software, you just ask. The AI understands what you want, talks to the CRM, and gives you the answer or takes the action — in seconds. This is the future of business software, and Webtrix24 is building it now.
| Task | Without MCP (Today) | With Webtrix24 MCP |
|---|---|---|
| Check expiring policies | Login → Reports → Filter → Export | “Show expiring policies this month” |
| Add a new lead | Login → Leads → Add New → Fill form | “Add lead: John, 9876543210, motor insurance” |
| Check idle inventory | Login → Inventory → Filter idle → Review | “Which laptops have been idle 10+ days?” |
| Send follow-up reminders | Login → Campaigns → Create → Send | “Send WhatsApp follow-up to cold leads” |
Which AI Assistants Will Work with Webtrix24 MCP?
The MCP protocol is an open standard, which means once Webtrix24 has an MCP server, it works with any MCP-compatible AI tool. The growing list includes:
- 🤖 Claude by Anthropic — The most capable AI assistant for business tasks, now available at claude.ai
- 💻 Cursor — AI-powered code and productivity tool
- 🔧 VS Code with Copilot — Microsoft’s developer AI tool
- 🌐 Any future MCP-compatible AI — The standard is being adopted rapidly across the industry
What Webtrix24 MCP Will Be Able to Do (Planned Features)
| Category | Actions via AI |
|---|---|
| Leads | Add lead, update status, assign to agent, list by filter |
| Customers | Search customer, view policy history, update contact details |
| Policies (Insurance) | Add policy, check expiry, list renewals due, update status |
| Inventory (Rental) | Check availability, view deployed assets, list idle inventory |
| Invoices | Generate invoice, check payment status, list overdue |
| Delivery Challans | Create challan, add items, share via WhatsApp |
| Campaigns | List upcoming campaigns, trigger manual campaign |
| Reports | Revenue summary, renewal rate, asset utilisation |
How We Are Building It: Technical Overview
For our technical readers and fellow developers — here’s how the Webtrix24 MCP Server is being built on our existing React, CodeIgniter 3, and MySQL stack.
Our current tech stack
- Frontend: React.js
- Backend: CodeIgniter 3 (PHP)
- Database: MySQL
- Communication: REST APIs between React and CI3
What MCP requires
An MCP server is a lightweight service that:
- Exposes a list of “tools” — named actions the AI can call (e.g. get_leads, add_policy, check_inventory)
- Receives requests from the AI with parameters (e.g. customer_name, date_range)
- Calls your existing REST APIs to execute those actions
- Returns structured responses the AI can understand and present to the user
The architecture — 3 layers
| Layer | Technology | Role |
|---|---|---|
| AI Assistant | Claude / any MCP client | User talks to AI, AI calls MCP tools |
| MCP Server (new) | Node.js + MCP SDK | Translates AI requests into API calls |
| Webtrix24 Backend | CodeIgniter 3 REST API + MySQL | Executes actions, returns data — no changes needed |
Step 1 — Build REST API endpoints in CodeIgniter 3
The first step is ensuring Webtrix24 has clean, authenticated REST API endpoints for every action the MCP server needs to perform. Most of these already exist since our React frontend calls them. We need to add token-based authentication (API key per user) and ensure consistent JSON responses.
Example CI3 endpoints needed:
GET /api/leads?status=new&limit=10
POST /api/leads/add
GET /api/policies?expiry_within_days=30
GET /api/inventory?status=idle
POST /api/invoices/generate
GET /api/customers/search?name=rahul
Step 2 — Build the MCP Server in Node.js
The MCP server is a separate Node.js service that uses Anthropic’s official MCP SDK. It defines tools — each tool maps to one of your CI3 API endpoints. Here’s a simplified example of what one tool looks like:
// Tool: get_expiring_policies
{
name: "get_expiring_policies",
description: "Get all insurance policies expiring within a given number of days",
inputSchema: {
type: "object",
properties: {
days: { type: "number", description: "Days until expiry" }
}
},
handler: async ({ days }) => {
const response = await fetch(
`https://app.webtrix24.com/api/policies?expiry_within_days=${days}`,
{ headers: { "Authorization": `Bearer ${apiKey}` } }
);
const data = await response.json();
return { policies: data.policies };
}
}
When a user asks Claude “show me policies expiring in 30 days,” Claude automatically calls this tool with days: 30, gets the data from your CI3 backend, and presents it in a readable format.
Step 3 — Add API key management in React frontend
Users need a way to generate and manage their MCP connection key. We’re adding a simple settings page in the Webtrix24 React dashboard where users can:
- Generate a personal API key for MCP connection
- Copy the MCP server URL and key to paste into their AI tool
- See which AI tools are currently connected
- Revoke access if needed
Step 4 — Security and permissions
Each API key is scoped to a specific user account in MySQL. The MCP server checks permissions on every request — an agent can only access their own data, while an admin can access team-wide data. All connections use HTTPS and JWT token validation through CI3’s existing auth middleware.
Step 5 — Deploy MCP server alongside existing infrastructure
The Node.js MCP server runs as a separate microservice on the same server infrastructure. It doesn’t replace or modify the CI3 backend — it simply sits alongside it as a new entry point that the AI uses, while your React frontend continues to work exactly as it does today.
Development Roadmap
| Phase | What We’re Building | Timeline |
|---|---|---|
| Phase 1 | Core REST API cleanup and API key system in CI3 | Month 1 |
| Phase 2 | MCP Server with read-only tools (leads, policies, inventory queries) | Month 2 |
| Phase 3 | Write tools (add lead, create challan, generate invoice) | Month 3 |
| Phase 4 | React settings page for MCP connection management | Month 3 |
| Phase 5 | Beta testing with select users | Month 4 |
| Phase 6 | Public launch — MCP Server available to all users | Month 5 |
Who Will Benefit Most from Webtrix24 MCP?
Insurance agents and brokers
Instead of logging into the CRM between every customer call, agents can ask Claude directly: “What is the renewal status of Amit Verma?” and get an instant answer — without switching apps or losing focus on the conversation.
Rental business managers
Operations managers can ask: “What’s our inventory utilisation this week?” or “List all overdue rentals with customer phone numbers” and get an instant, actionable report without running queries or exporting data.
Business owners reviewing performance
Instead of pulling monthly reports manually, owners can ask their AI: “How many new leads did we add last month?” or “What was our renewal rate in April?” — and get instant data-driven answers.
Be the First to Try Webtrix24 MCP
We’re building this feature now and looking for early beta users — insurance agents, rental businesses, and CRM power users who want to be the first to experience AI-powered CRM management.
If you’re interested in joining the beta, register your interest below and we’ll reach out when it’s ready for testing. Your feedback will directly shape what we build.
Frequently Asked Questions
What is MCP and how does it work with Webtrix24?
MCP (Model Context Protocol) is an open standard that lets AI assistants like Claude connect directly to business software. The Webtrix24 MCP Server will let you manage your CRM — add leads, check renewals, view inventory — just by talking to an AI assistant, without logging into the CRM dashboard.
Do I need to be technical to use the Webtrix24 MCP Server?
No. Once set up, using MCP is as simple as having a conversation with an AI assistant. The technical setup involves pasting your Webtrix24 API key into your AI tool — a one-time, two-minute process. Everything else is just natural language.
Which AI tools will work with Webtrix24 MCP?
The initial launch will support Claude by Anthropic (claude.ai). Additional MCP-compatible tools including Cursor and others will be added in subsequent releases as the MCP standard becomes more widely adopted.
Is my data safe when using MCP?
Yes. Every MCP connection uses a personal API key scoped to your account. All data transfer is encrypted via HTTPS. You can revoke MCP access at any time from your Webtrix24 settings page. The AI assistant never stores your CRM data — it only reads and writes in real time.
When will the Webtrix24 MCP Server be available?
We’re targeting a beta release in approximately 3–4 months, with a public launch to all users shortly after. Register your interest via WhatsApp to be notified when beta access opens.
Will MCP be available on the free plan?
Yes — basic MCP read access (queries and reports) will be available on all plans including the free plan. Advanced write actions (creating records, sending campaigns) will be part of paid plans.