MCP Guide

Integrate AdCP with Model Context Protocol (MCP). Transport-specific guide for tool calls, context management, and wire format.

Transport-specific guide for integrating AdCP using the Model Context Protocol. For task handling, status management, and workflow patterns, see Core Concepts.

Testing AdCP via MCP

You can test AdCP tasks using the reference implementation at test-agent.adcontextprotocol.org. This endpoint implements all AdCP tasks as MCP tools and is useful for development and integration testing.

Tool Call Patterns

Basic Tool Invocation

// Standard MCP tool call
const response = await mcp.call('get_products', {
  brand_manifest: {
    name: "Premium Pet Foods",
    url: "https://premiumpetfoods.com"
  },
  brief: "Video campaign for pet owners"
});

// All responses include status field (AdCP 1.6.0+)
console.log(response.status);   // "completed" | "input-required" | "working" | etc.
console.log(response.message);  // Human-readable summary

Tool Call with Filters

// Structured parameters
const response = await mcp.call('get_products', {
  brand_manifest: {
    name: "BetNow",
    url: "https://betnow.com"
  },
  brief: "Sports betting app for March Madness",
  filters: {
    format_types: ["video"],
    delivery_type: "guaranteed",
    max_cpm: 50
  }
});

Tool Call with Application-Level Context

// Pass opaque application-level context; agents must carry it back
const response = await mcp.call('build_creative', {
  target_format_id: { agent_url: 'https://creative.agent', id: 'premium_bespoke_display' },
  creative_manifest: { /* ... */ },
  context: { ui: 'buyer_dashboard', session: '123' }
});

// Response includes the same context at the top level
console.log(response.context); // { ui: 'buyer_dashboard', session: '123' }

MCP Response Format

New in AdCP 1.6.0: All responses include unified status field.

MCP responses use a flat structure where task-specific fields are at the top level alongside protocol fields:

{
  "status": "completed",           // Unified status (see Core Concepts)
  "message": "Found 5 products",   // Human-readable summary
  "context_id": "ctx-abc123",      // MCP session continuity
  "context": { "ui": "buyer_dashboard" }, // Application-level context echoed back
  "products": [...],               // Task-specific data (flat, not nested)
  "errors": [...]                  // Task-level errors/warnings
}

MCP-Specific Fields

• context_id: Session identifier that you must manually manage
• context: Opaque initiator-provided metadata echoed by agents
• status: Same values as A2A protocol for consistency
• Task-specific fields (e.g., products, media_buy_id, creatives) are at the top level, not wrapped in a data object

Status Handling: See Core Concepts for complete status handling patterns.

Available Tools

All AdCP tasks are available as MCP tools:

Media Buy Tools

await mcp.call('get_products', {...});           // Discover inventory
await mcp.call('list_creative_formats', {...});  // Get format specs
await mcp.call('create_media_buy', {...});       // Create campaigns  
await mcp.call('update_media_buy', {...});       // Modify campaigns
await mcp.call('sync_creatives', {...});         // Manage creative assets
await mcp.call('get_media_buy_delivery', {...}); // Performance metrics
await mcp.call('list_authorized_properties', {...}); // Available properties
await mcp.call('provide_performance_feedback', {...}); // Share outcomes

Task Management Tools

await mcp.call('tasks/list', {...});          // List and filter async tasks
await mcp.call('tasks/get', {...});           // Poll specific task status

Signals Tools

await mcp.call('get_signals', {...});      // Discover audience signals
await mcp.call('activate_signal', {...});  // Deploy signals to platforms

Task Parameters: See individual task documentation in Media Buy and Signals sections.

Task Management: For comprehensive guidance on tracking async operations, polling patterns, and webhook integration, see Task Management.

Context Management (MCP-Specific)

Critical: MCP requires manual context management. You must pass context_id to maintain conversation state.

Context Session Pattern

class McpAdcpSession {
  constructor(mcpClient) {
    this.mcp = mcpClient;
    this.contextId = null;
    this.activeTasks = new Map(); // Track async operations
  }
  
  async call(tool, params, options = {}) {
    // Build request with protocol-level fields
    const request = {
      tool: tool,
      arguments: params
    };
    
    // Include context from previous calls
    if (this.contextId) {
      request.context_id = this.contextId;
    }
    
    // Include webhook configuration (protocol-level, A2A-compatible)
    if (options.push_notification_config) {
      request.push_notification_config = options.push_notification_config;
    }
    
    const response = await this.mcp.call(request);
    
    // Save context for next call
    this.contextId = response.context_id;
    
    // Track async operations
    if (response.task_id) {
      this.activeTasks.set(response.task_id, {
        tool,
        params,
        startTime: new Date(),
        status: response.status
      });
    }
    
    return response;
  }
  
  reset() {
    this.contextId = null;
    this.activeTasks.clear();
  }
  
  // Poll specific task
  async pollTask(taskId, includeResult = false) {
    return this.call('tasks/get', { 
      task_id: taskId, 
      include_result: includeResult 
    });
  }
  
  // List pending tasks
  async listPendingTasks() {
    return this.call('tasks/list', {
      filters: {
        statuses: ["submitted", "working", "input-required"]
      }
    });
  }
  
  // State reconciliation helper
  async reconcileState() {
    const pending = await this.listPendingTasks();
    const serverTasks = new Set(pending.tasks.map(t => t.task_id));
    const clientTasks = new Set(this.activeTasks.keys());
    
    return {
      missing_from_client: [...serverTasks].filter(id => !clientTasks.has(id)),
      missing_from_server: [...clientTasks].filter(id => !serverTasks.has(id)),
      total_pending: pending.tasks.length
    };
  }
}

Usage Examples

Basic Session with Context

const session = new McpAdcpSession(mcp);

// First call - no context needed
const products = await session.call('get_products', {
  brief: "Sports campaign"
});

// Follow-up - context automatically included
const refined = await session.call('get_products', {
  brief: "Focus on premium CTV"
});
// Session remembers previous interaction

Async Operations with Webhooks

MCP doesn’t define push notifications. AdCP fills this gap by specifying the webhook configuration (pushNotificationConfig) and payload format (mcp-webhook-payload.json). When you configure a webhook, the server will POST task updates to your URL instead of requiring you to poll.

Webhook Envelope: mcp-webhook-payload.json

Best Practice: URL-Based Routing

Recommended: Encode routing information (task_type, operation_id) in the webhook URL, not the payload.

Why this approach?

• ✅ Industry standard pattern - Widely adopted for webhook routing across major APIs
• ✅ Separation of concerns - URLs handle routing, payloads contain data
• ✅ Protocol-agnostic - Same pattern works for MCP, A2A, REST, future protocols
• ✅ Simpler handlers - Route with URL framework, not payload parsing

URL Pattern Options:

// Option 1: Path parameters (recommended)
url: `https://buyer.com/webhooks/adcp/${taskType}/${operationId}`
// Example: /webhooks/adcp/create_media_buy/op_nike_q1_2025

// Option 2: Query parameters
url: `https://buyer.com/webhooks/adcp?task=${taskType}&op=${operationId}`

// Option 3: Subdomain routing
url: `https://${taskType}.webhooks.buyer.com/${operationId}`

Example Configuration:

const operationId = "op_nike_q1_2025";
const taskType = "create_media_buy";

// Configure webhook with routing in URL
const response = await session.call('create_media_buy',
  {
    buyer_ref: "nike_q1_2025",
    packages: [...],
    budget: { total: 150000, currency: "USD" }
  },
  {
    pushNotificationConfig: {
      url: `https://buyer.com/webhooks/adcp/${taskType}/${operationId}`,
      authentication: {
        schemes: ["HMAC-SHA256"],  // or ["bearer"] for simple auth
        credentials: "shared_secret_32_chars"
      }
    }
  }
);

if (response.status === 'submitted') {
  console.log(`Task ${response.task_id} submitted for long-running execution`);
  // Server will POST status updates to your webhook URL
} else if (response.status === 'completed') {
  console.log(`Media buy created: ${response.media_buy_id}`);
}

Webhook POST format:

{
  "task_id": "task_456",
  "status": "completed",
  "timestamp": "2025-01-22T10:30:00Z",
  "result": {
    "media_buy_id": "mb_12345",
    "buyer_ref": "nike_q1_2025",
    "packages": [...]
  }
}

Note: This example follows the recommended URL-based routing pattern where task_type and operation_id are passed in the URL (e.g., /webhooks/adcp/create_media_buy/op_456). While the schema still supports these fields in the payload for backward compatibility, they are deprecated.

The result field contains the AdCP data payload. For completed/failed statuses, this is the full task response (e.g., create-media-buy-response.json). For other statuses, use the status-specific schemas (e.g., create-media-buy-async-response-working.json).

MCP Webhook Envelope Fields

The mcp-webhook-payload.json envelope includes:

Required fields:

• task_id — Unique task identifier for correlation
• status — Current task status (completed, failed, working, input-required, etc.)
• timestamp — ISO 8601 timestamp when webhook was generated

Optional fields:

• domain — AdCP domain (“media-buy” or “signals”)
• context_id — Conversation/session identifier
• message — Human-readable context about the status change

Deprecated fields (supported but not recommended):

• task_type — Task name (e.g., “create_media_buy”, “sync_creatives”) - ⚠️ Deprecated: See URL-Based Routing
• operation_id — Correlates a sequence of updates for the same operation - ⚠️ Deprecated: See URL-Based Routing

Data field:

• result — Task-specific AdCP payload (see Data Schema Validation below)

Webhook Trigger Rules

Webhooks are sent when all of these conditions are met:

1. Task type supports async (e.g., create_media_buy, sync_creatives, get_products)
2. pushNotificationConfig is provided in the request
3. Task runs asynchronously — initial response is working or submitted

If the initial response is already terminal (completed, failed, rejected), no webhook is sent—you already have the result.

Status changes that trigger webhooks:

• working → Progress update (task actively processing)
• input-required → Human input needed
• completed → Final result available
• failed → Error details

Data Schema Validation

The result field in MCP webhooks uses status-specific schemas:

Status	Schema	Contents
completed	[task]-response.json	Full task response (success branch)
failed	[task]-response.json	Full task response (error branch)
working	[task]-async-response-working.json	Progress info (percentage, step)
input-required	[task]-async-response-input-required.json	Requirements, approval data
submitted	[task]-async-response-submitted.json	Acknowledgment (usually minimal)

Schema reference: async-response-data.json

Webhook Handler Example

const express = require('express');
const app = express();

app.post('/webhooks/adcp/:task_type/:agent_id/:operation_id', async (req, res) => {
  const { task_type, agent_id, operation_id } = req.params;
  const webhook = req.body;

  // Verify webhook authenticity (HMAC-SHA256 example)
  const signature = req.headers['x-adcp-signature'];
  const timestamp = req.headers['x-adcp-timestamp'];
  if (!verifySignature(webhook, signature, timestamp)) {
    return res.status(401).json({ error: 'Invalid signature' });
  }

  // Handle status changes
  switch (webhook.status) {
    case 'input-required':
      // Alert human that input is needed
      await notifyHuman({
        operation_id,
        message: webhook.message,
        context_id: webhook.context_id,
        data: webhook.result
      });
      break;

    case 'completed':
      // Process the completed operation
      if (task_type === 'create_media_buy') {
        await handleMediaBuyCreated({
          media_buy_id: webhook.result.media_buy_id,
          buyer_ref: webhook.result.buyer_ref,
          packages: webhook.result.packages
        });
      }
      break;

    case 'failed':
      // Handle failure
      await handleOperationFailed({
        operation_id,
        error: webhook.result?.errors,
        message: webhook.message
      });
      break;

    case 'working':
      // Update progress UI
      await updateProgress({
        operation_id,
        percentage: webhook.result?.percentage,
        message: webhook.message
      });
      break;

    case 'canceled':
      await handleOperationCanceled(operation_id, webhook.message);
      break;
  }

  // Always return 200 for successful processing
  res.status(200).json({ status: 'processed' });
});

function verifySignature(payload, signature, timestamp) {
  const crypto = require('crypto');
  const expectedSig = crypto
    .createHmac('sha256', process.env.WEBHOOK_SECRET)
    .update(timestamp + JSON.stringify(payload))
    .digest('hex');
  return signature === `sha256=${expectedSig}`;
}

Task Management and Polling

// Check status of specific task
const taskStatus = await session.pollTask('task_456', true);
if (taskStatus.status === 'completed') {
  console.log('Result:', taskStatus.result);
}

// State reconciliation
const reconciliation = await session.reconcileState();
if (reconciliation.missing_from_client.length > 0) {
  console.log('Found orphaned tasks:', reconciliation.missing_from_client);
  // Start tracking these tasks
}

// List all pending operations
const pending = await session.listPendingTasks();
console.log(`${pending.tasks.length} operations in progress`);

Context Expiration Handling

async function handleContextExpiration(session, tool, params) {
  try {
    return await session.call(tool, params);
  } catch (error) {
    if (error.message?.includes('context not found')) {
      // Context expired - start fresh
      session.reset();
      return session.call(tool, params);
    }
    throw error;
  }
}

Key Difference: Unlike A2A which manages context automatically, MCP requires explicit context_id management.

Handling Async Operations

When a task returns working or submitted status, you have two options for receiving updates:

Approach	Best For	Trade-offs
Polling	Simple integrations, short tasks	Easy to implement, but inefficient for long waits
Webhooks	Production systems, long-running tasks	More efficient, but requires a public endpoint

Option 1: Polling

Use tasks/get to check task status periodically:

async function waitForCompletion(session, initialResponse) {
  if (!initialResponse.task_id) {
    return initialResponse; // Already completed
  }
  
  // Poll more frequently for 'working' (will finish soon)
  // Poll less frequently for 'submitted' (may take hours)
  let pollInterval = initialResponse.status === 'working' ? 5000 : 30000;
  
  while (true) {
    const response = await session.pollTask(initialResponse.task_id, true);
    
    if (['completed', 'failed', 'canceled'].includes(response.status)) {
      return response;
    }
    
    if (response.status === 'input-required') {
      const input = await promptUser(response.message);
      return session.call('create_media_buy', {
        context_id: response.context_id,
        additional_info: input
      });
    }
    
    pollInterval = response.status === 'working' ? 5000 : 30000;
    await new Promise(resolve => setTimeout(resolve, pollInterval));
  }
}

Option 2: Webhooks

Configure a webhook URL and the server will POST updates to you directly. This is more efficient for long-running tasks since you don’t need to keep polling.

const response = await session.call('create_media_buy',
  {
    buyer_ref: "nike_q1_2025",
    packages: [...],
    budget: { total: 150000, currency: "USD" }
  },
  {
    pushNotificationConfig: {
      url: "https://buyer.com/webhooks/adcp",
      authentication: {
        schemes: ["HMAC-SHA256"],
        credentials: "shared_secret_32_chars"
      }
    }
  }
);

// If status is 'submitted', the server will POST updates to your webhook
// You don't need to poll - just wait for the webhook

See Task Management for webhook payload formats and handling examples.

Handling Different Statuses

const initial = await session.call('create_media_buy', {
  buyer_ref: "nike_q1_2025",
  packages: [...],
  budget: { total: 100000, currency: "USD" }
});

switch (initial.status) {
  case 'completed':
    // Done immediately - no async handling needed
    console.log('Created:', initial.media_buy_id);
    break;
    
  case 'working':
    // Will finish within ~2 minutes - poll or wait
    console.log('Processing...');
    const final = await waitForCompletion(session, initial);
    console.log('Created:', final.result.media_buy_id);
    break;
    
  case 'submitted':
    // Long-running (hours/days) - use webhooks or poll infrequently
    console.log(`Task ${initial.task_id} queued for approval`);
    // Webhook will notify when complete, or poll manually
    break;
    
  case 'input-required':
    // Blocked on user input
    console.log('Need more info:', initial.message);
    break;
}

Integration Example

// Initialize MCP session with context management
const session = new McpAdcpSession(mcp);

// Use unified status handling (see Core Concepts)
async function handleAdcpCall(tool, params, options = {}) {
  const response = await session.call(tool, params, options);
  
  switch (response.status) {
    case 'input-required':
      // Handle clarification (see Core Concepts for patterns)
      const input = await promptUser(response.message);
      return session.call(tool, { ...params, additional_info: input });
      
    case 'working':
      // Handle short async operations 
      return waitForCompletion(session, response);
      
    case 'submitted':
      // Handle long async operations
      if (options.webhook_url) {
        console.log(`Task ${response.task_id} submitted, webhook will notify`);
        return { pending: true, task_id: response.task_id };
      } else {
        console.log(`Task ${response.task_id} submitted, polling...`);
        return waitForCompletion(session, response);
      }
      
    case 'completed':
      return response; // Task-specific fields are at the top level
      
    case 'failed':
      throw new Error(response.message);
  }
}

// Example usage
const products = await handleAdcpCall('get_products', {
  brief: "CTV campaign for luxury cars"
});

MCP-Specific Considerations

Tool Discovery

// List available AdCP tools
const tools = await mcp.listTools();
const adcpTools = tools.filter(t => t.name.startsWith('adcp_') || 
  ['get_products', 'create_media_buy'].includes(t.name));

AdCP Extension (Future)

Status: MCP server cards are expected in a future MCP release. When available, AdCP servers will include the AdCP extension.

{
  "extensions": {
    "adcp": {
      "adcp_version": "2.4.0",
      "protocols_supported": ["media_buy", "creative", "signals"]
    }
  }
}

This will allow clients to programmatically discover which AdCP version and protocol domains an MCP server implements. See the AdCP extension schema for specification details.

Parameter Validation

// MCP provides tool schemas for validation
const toolSchema = await mcp.getToolSchema('get_products');
// Use schema to validate parameters before calling

Error Handling

try {
  const response = await session.call('get_products', params);
} catch (mcpError) {
  // MCP transport errors (connection, auth, etc.)
  console.error('MCP Error:', mcpError);
} 

// AdCP task errors come in response.status === 'failed'

Best Practices

1. Use session wrapper for automatic context management
2. Check status field before processing response data
3. Handle context expiration gracefully with retries
4. Reference Core Concepts for status handling patterns
5. Validate parameters using MCP tool schemas when available

Next Steps

• Core Concepts: Read Core Concepts for status handling and workflows
• Task Reference: See Media Buy Tasks and Signals
• Protocol Comparison: Compare with A2A integration
• Examples: Find complete workflow examples in Core Concepts

For status handling, async operations, and clarification patterns, see Core Concepts - this guide focuses on MCP transport specifics only.