Skip to main content
Model Context Protocol

Connect any AI agent to your dispatch workflow.

DispatchNode exposes a full MCP server so Claude, GPT, or any AI agent can book jobs, check availability, pull transcripts, and monitor health — all through your existing org credentials.

What is MCP?

The USB-C for AI agents.

The Model Context Protocol (MCP) is an open standard that lets AI agents connect to external tools and data sources through a single, consistent interface. Instead of building custom integrations for every AI platform, you expose your capabilities once as MCP tools — and every compliant agent can use them.

Works with Claude, GPT, Gemini, and any MCP client
Typed inputs/outputs — no hallucinated API calls
Bearer token auth — your data stays scoped to your org
Real-time health signals — agents know before they call
mcp-client.ts
// Connect to DispatchNode MCP
const response = await fetch(
  "https://dispatchnode.com/api/mcp",
  {
    method: "POST",
    headers: {
      "Content-Type": "application/json",
      "Authorization": "Bearer dnk_your_key"
    },
    body: JSON.stringify({
      jsonrpc: "2.0",
      id: 1,
      method: "tools/call",
      params: {
        name: "book_job",
        arguments: {
          customerName: "John Smith",
          customerPhone: "+15551234567",
          serviceAddress: "123 Main St",
          scheduledStart: "2026-04-15T09:00"
        }
      }
    })
  }
);

Available Tools

6 tools. Full dispatch control.

book_job

Book a new service job with customer details, scheduling, and service type.

Inputs

customerName, customerPhone, serviceAddress, scheduledStart

check_availability

Query open time slots across your calendar with transit buffer awareness.

Inputs

startDate, endDate, serviceType?

list_jobs

List active jobs with status filter, pagination, and customer context.

Inputs

status?, limit?, offset?

get_pricing

Retrieve service pricing, deposit requirements, and plan metadata.

Inputs

serviceType?

get_call_transcript

Pull transcript and metadata for any AI voice call by ID.

Inputs

callId

get_health

System health: DB status, circuit breaker states, latency metrics.

Inputs

(none)

Authentication

Two tiers. Zero friction.

Customer-facing AI agents can book jobs, check availability, and get pricing with no API key — just the org slug from your site. Business owners get full admin access via a secure API key.

Customer Tier
No API key needed. AI agents pass X-Org-Slug header to book jobs, check slots, and get pricing.
Operator Tier
dnk_ API key for full access — list jobs, view transcripts, check system health.
Rate Limited
Customer tier: 30 requests/min, 10 bookings/hour per IP. Operator tier: unlimited.
Tenant Isolated
Every call is scoped to one organization. No cross-tenant access, ever.
Customer Tier — No Key
X-Org-Slug: eventrestroomrentals

→ book_job
→ check_availability
→ get_pricing
Operator Tier — Full Access
Authorization: Bearer dnk_a1b2c3d4...

→ All 6 tools
→ list_jobs, get_call_transcript
→ get_health

Beyond Tool Calling

5 protocol layers. All implemented.

Tool calling handles invocation. We handle everything that makes selection, billing, and repeat use safe.

Typed I/O Schemas

Every tool has a full JSON Schema definition for inputs and outputs. Zod-validated server-side. No guesswork.

Async Job Lifecycle

Jobs move through a state machine (LEAD → SCHEDULED → IN_PROGRESS → COMPLETED). Dead-letter queue captures failures.

Auth + Pricing Metadata

Org-scoped API keys with SHA-256 hashing. Pricing metadata attached to every service type tool response.

Retries & Idempotency

Exponential backoff with jitter for transient failures. Redis-backed idempotency keys prevent duplicate processing.

Health & Reputation

Circuit breaker (CLOSED/OPEN/HALF_OPEN) per external service. SLI metrics for latency, error rates, and throughput.

Quick Start

3 steps. Up and running.

01

Generate an API Key

Go to Dashboard → Settings → API Keys. Create a new key and copy it — it's only shown once.

// Your key looks like:
dnk_a1b2c3d4e5f6g7h8i9j0k1l2m3n4o5p6q7r8
02

Initialize the Connection

Send an initialize request to establish the MCP session.

POST /api/mcp
Authorization: Bearer dnk_your_key

{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "initialize",
  "params": {}
}
03

Call a Tool

Use tools/call with the tool name and typed arguments.

{
  "jsonrpc": "2.0",
  "id": 2,
  "method": "tools/call",
  "params": {
    "name": "check_availability",
    "arguments": {
      "startDate": "2026-04-15",
      "endDate": "2026-04-16"
    }
  }
}

Ready to integrate. Start now.

Sign up, generate your API key, and connect your first AI agent to DispatchNode in minutes.