> ## Documentation Index > Fetch the complete documentation index at: https://help.bodharouteplanner.com/llms.txt > Use this file to discover all available pages before exploring further. # AI-Powered Fleet Intelligence > Talk to your fleet data. Connect any AI assistant to Bodha Route Planner in 2 minutes — zero code, zero install. # Talk to your fleet data What if you could ask your fleet a question — and get an answer? > *"How did my drivers perform this week?"* > *"Which stops failed yesterday and why?"* > *"Compare this month to last month."* > *"Are there any routes still in progress?"* No dashboards to navigate. No reports to export. No spreadsheets to cross-reference. Just **ask**. Bodha Route Planner's MCP server connects your delivery data directly to AI assistants like **Claude**, **Cursor**, **Windsurf**, and any MCP-compatible client. Your AI assistant becomes a fleet analyst that knows every route, every stop, every driver — and can answer any question in seconds. *** ## What is MCP? **MCP (Model Context Protocol)** is the open standard that lets AI assistants call external tools and data sources. It's supported by every major AI client — Claude Desktop, Cursor, Windsurf, VS Code Copilot, and more. Think of it this way: | Without MCP | With MCP | | - | - | | Open dashboard → Analytics → pick date range → read charts → export → cross-reference in Excel | *"Give me this week's summary"* → done | | Navigate to routes → filter by failed → click each one → read the reason | *"Which stops failed today and what were the reasons?"* → instant table with analysis | | Check 5 driver profiles individually → compare numbers mentally | *"Compare all my drivers' performance this month"* → ranked comparison with insights | *** ## How it works **1. Generate an API key** Create a scoped API key in your Bodha Route Planner dashboard. Takes 30 seconds. **2. Paste one config block** Add a small JSON config to your AI assistant. We give you the exact text to copy. **3. Start asking questions** Open your AI assistant and ask anything about your routes, drivers, stops, or orders. The assistant calls Bodha Route Planner behind the scenes and gives you the answer in plain language. **No npm install. No Docker. No local server. No code.** It's a hosted service that works immediately. *** ## What you can ask The MCP server exposes **14 tools** that cover your entire operation: > **Routes** > > List today's routes, check status, see which ones are complete or still running. > **Stops** > > Search for deliveries by name or address, check what failed and why, see completion details. > **Drivers** > > See who's on the road, who's available, compare driver performance over any period. > **Orders** > > Check your orders inbox, see what's pending, what's been assigned, what's done. > **Analytics** > > Get instant summaries — completion rates, on-time delivery, distance, trends. Compare any two periods. > **Depots** > > View your depot locations and details. *** ## Real conversations Here's what it looks like in practice: **Morning check-in:** > **You:** "What's happening today?" > > **AI:** "You have 8 routes today with 247 total stops across 8 drivers. 3 routes are in progress (112 stops completed so far), 5 are dispatched and waiting to start. Current completion rate is 89%. No failed stops yet today." **End-of-day review:** > **You:** "How did we do this week vs last week?" > > **AI:** "This week: 42 routes, 1,340 stops, 94.2% completion rate, average 31.9 stops per route. Last week: 38 routes, 1,180 stops, 91.8% completion rate, 31.1 stops per route. That's a 2.4% improvement in completion and 13.6% more stops delivered. Your fleet is trending up." **Investigating a problem:** > **You:** "Which stops failed yesterday and what were the top reasons?" > > **AI:** "14 stops failed yesterday across 6 routes. Top reasons: Not home (7), Refused delivery (3), Wrong address (2), Access issue (1), Damaged package (1). The 'not home' failures were concentrated in the afternoon routes — you might consider tighter time windows for residential stops." *** ## Security first | | | | - | - | | **Tenant isolation** | Your data is yours. Every query is filtered by your organization. You can never see another tenant's data — by design, not by policy. | | **Encrypted keys** | API keys are stored as bcrypt hashes. We never store the plaintext. | | **Scoped access** | Each key has specific scopes — you control exactly what the AI can see. | | **Read-only** | The MCP server can read your data but **cannot modify anything** — no creating, editing, or deleting. | | **Instant revocation** | Revoke a key and it stops working on the very next request. | | **Audit trail** | Every tool call is logged — what was asked, when, which key, how long it took. | *** ## Pricing | Plan | MCP access | | - | - | | **Starter** | — | | **Pro** | 1 API key, all read scopes | | **Enterprise** | Unlimited keys, all scopes, full audit log | **No per-call charges.** MCP is included in your plan. The more you use it, the more value you get — we don't meter it. *** ## Get started > **[Setup guide](https://help.bodharouteplanner.com/mcp/setup.md)** > > Generate your key and connect your AI assistant in 2 minutes. > **[Connect Claude Desktop](https://help.bodharouteplanner.com/mcp/connect-claude.md)** > > Step-by-step for Anthropic's Claude Desktop app. > **[Connect Cursor](https://help.bodharouteplanner.com/mcp/connect-cursor.md)** > > Step-by-step for the Cursor AI code editor. > **[All tools reference](https://help.bodharouteplanner.com/mcp/tools.md)** > > Every tool, every parameter, every return field — documented. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.