Prebid Sales Agent

A reference implementation of the Ad Context Protocol (AdCP) sales agent, enabling AI agents to buy advertising inventory through a standardized MCP interface. This codebase is maintained under Prebid.org as the Prebid Sales Agent reference implementation.

What is this?

The Prebid Sales Agent is a server that:

• Exposes advertising inventory to AI agents via MCP (Model Context Protocol) and A2A (Agent-to-Agent)
• Integrates with ad servers like Google Ad Manager
• Provides an admin interface for managing inventory and campaigns
• Handles the full campaign lifecycle from discovery to reporting

Choose Your Path

I want to...	Start here
Deploy my own sales agent (publisher)	Quickstart Guide
Evaluate or develop locally	Quick Start below
Run a multi-tenant platform	Deployment Guide

---

Quick Start (Evaluation)

Try the sales agent locally:

# Clone and start
git clone https://github.com/prebid/salesagent.git
cd salesagent
docker compose up -d

# Test the MCP interface
uvx adcp http://localhost:8000/mcp/ --auth test-token list_tools
uvx adcp http://localhost:8000/mcp/ --auth test-token get_products '{"brief":"video"}'

Access services at http://localhost:8000:

• Admin UI: /admin or just click "Log in to Dashboard" (test credentials: test123)
• MCP Server: /mcp/
• A2A Server: /a2a

For production deployment, see the Quickstart Guide.

---

Publisher Deployment

Publishers deploy their own sales agent. Choose based on your needs:

Platform	Time	Difficulty	Guide
Docker (local/on-prem)	2 min	Easy	quickstart.md
Fly.io (cloud)	10-15 min	Medium	fly.md
Google Cloud Run	15-20 min	Medium	gcp.md

Docker is the fastest - it bundles PostgreSQL and just works. Cloud platforms require separate database setup.

After Deployment

Configure via the Admin UI:

1. Configure your ad server (Settings → Adapters)
2. Set up products that match your GAM line items
3. Add advertisers who will use the MCP API
4. Set your custom domain (Settings → General)

---

Development Setup

git clone https://github.com/prebid/salesagent.git
cd salesagent
make setup    # One command: installs deps, starts Docker, verifies health

See the Getting Started guide for prerequisites, manual setup steps, testing workflows, and common operations.

---

Google Ad Manager Setup

For GAM integration, choose your authentication method:

Service Account (Recommended for Production):

• No OAuth credentials needed
• Configure service account JSON in Admin UI
• See GAM Adapter Guide for setup

OAuth (Development/Testing):

1. Create OAuth credentials at Google Cloud Console
2. Add to .env:

   GAM_OAUTH_CLIENT_ID=your-client-id.apps.googleusercontent.com
   GAM_OAUTH_CLIENT_SECRET=your-client-secret
   

3. Configure in Admin UI: Settings → Adapters → Google Ad Manager

---

Using with Claude Desktop

Add to your Claude config (~/Library/Application Support/Claude/claude_desktop_config.json):

{
  "mcpServers": {
    "adcp": {
      "command": "uvx",
      "args": ["mcp-remote", "http://localhost:8000/mcp/", "--header", "x-adcp-auth: YOUR_TOKEN"]
    }
  }
}

Get your token from Admin UI → Advertisers → (select advertiser) → API Token.

---

Troubleshooting

Container won't start?

docker compose logs adcp-server | head -50

GAM OAuth error?

• Verify GAM_OAUTH_CLIENT_ID and GAM_OAUTH_CLIENT_SECRET in .env
• Restart: docker compose restart

More help: Troubleshooting Guide

Documentation

Deployment Guides

• Quickstart - Docker deployment (2 min)
• Fly.io - Cloud deployment (10-15 min)
• Google Cloud Run - GCP deployment (15-20 min)
• Single-Tenant - Single publisher deployment
• Multi-Tenant - Platform deployment

Reference

• Development Guide - Local development and contributing
• Architecture - System design and database schema
• Troubleshooting Guide - Monitoring and debugging

Key Features

For AI Agents

• Product Discovery - Natural language search for advertising products
• Campaign Creation - Automated media buying with targeting
• Creative Management - Upload and approval workflows
• Performance Monitoring - Real-time campaign metrics

For Publishers

• Multi-Tenant System - Isolated data per publisher
• Adapter Pattern - Support for multiple ad servers
• Real-time Dashboard - Live activity feed with Server-Sent Events (SSE)
• Workflow Management - Unified system for human-in-the-loop approvals
• Operations Monitoring - Track all media buys, workflows, and system activities
• Admin Interface - Web UI with Google OAuth
• Audit Logging - Complete operational history

For Developers

• MCP Protocol - Standard interface for AI agents
• A2A Protocol - Agent-to-Agent communication via JSON-RPC 2.0
• REST API - Programmatic tenant management
• Docker Deployment - Easy local and production setup
• Comprehensive Testing - Unit, integration, and E2E tests

Protocol Support

MCP (Model Context Protocol)

The primary interface for AI agents to interact with the Prebid Sales Agent. Uses FastMCP with HTTP/SSE transport.

A2A (Agent-to-Agent Protocol)

JSON-RPC 2.0 compliant server for agent-to-agent communication:

• Endpoint: /a2a (also available at port 8091)
• Discovery: /.well-known/agent.json
• Authentication: Bearer tokens via Authorization header
• Library: Built with standard python-a2a library

Testing Backend

The mock server provides comprehensive AdCP testing capabilities for developers:

Testing Headers Support

• X-Dry-Run: Test operations without real execution
• X-Mock-Time: Control time for deterministic testing
• X-Jump-To-Event: Skip to specific campaign events
• X-Test-Session-ID: Isolate parallel test sessions
• X-Auto-Advance: Automatic event progression
• X-Force-Error: Simulate error conditions

Response Headers

• X-Next-Event: Next expected campaign event
• X-Next-Event-Time: Timestamp for next event
• X-Simulated-Spend: Current campaign spend simulation

Testing Features

• Campaign Lifecycle Simulation: Complete event progression (creation → completion)
• Error Scenario Testing: Budget exceeded, delivery issues, platform errors
• Time Simulation: Fast-forward campaigns for testing
• Session Isolation: Parallel test execution without conflicts
• Production Safety: Zero real spend during testing

# Example: Test with time simulation
headers = {
    "x-adcp-auth": "your_token",
    "X-Dry-Run": "true",
    "X-Mock-Time": "2025-02-15T12:00:00Z",
    "X-Test-Session-ID": "test-123"
}

# Use with any MCP client for safe testing

See examples/mock_server_testing_demo.py for complete testing examples.

Using the MCP Client

from fastmcp.client import Client
from fastmcp.client.transports import StreamableHttpTransport

# Connect to server
headers = {"x-adcp-auth": "your_token"}
transport = StreamableHttpTransport(
    url="http://localhost:8000/mcp/",
    headers=headers
)
client = Client(transport=transport)

# Discover products
async with client:
    products = await client.tools.get_products(
        brief="video ads for sports content"
    )

    # Create media buy
    result = await client.tools.create_media_buy(
        product_ids=["ctv_sports"],
        total_budget=50000,
        flight_start_date="2025-02-01",
        flight_end_date="2025-02-28"
    )

Project Structure

salesagent/
├── src/                    # Source code
│   ├── core/              # Core MCP server components
│   │   ├── main.py        # MCP server implementation
│   │   ├── schemas.py     # API schemas and data models
│   │   ├── config_loader.py  # Configuration management
│   │   ├── audit_logger.py   # Security and audit logging
│   │   └── database/      # Database layer
│   │       ├── models.py  # SQLAlchemy models
│   │       ├── database.py # Database initialization
│   │       └── database_session.py # Session management
│   ├── services/          # Business logic services
│   │   ├── ai_product_service.py # AI product management
│   │   ├── targeting_capabilities.py # Targeting system
│   │   └── gam_inventory_service.py # GAM integration
│   ├── adapters/          # Ad server integrations
│   │   ├── base.py        # Base adapter interface
│   │   ├── google_ad_manager.py # GAM adapter
│   │   └── mock_ad_server.py # Mock adapter
│   └── admin/             # Admin UI (Flask)
│       ├── app.py         # Flask application
│       ├── blueprints/    # Flask blueprints
│       │   ├── tenants.py # Tenant dashboard
│       │   ├── tasks.py   # Task management (DEPRECATED - see workflow system)
│       │   └── activity_stream.py # Real-time activity feed
│       └── server.py      # Admin server
├── scripts/               # Utility scripts
│   ├── setup/            # Setup and initialization
│   ├── dev/              # Development tools
│   ├── ops/              # Operations scripts
│   └── deploy/           # Deployment scripts
├── tests/                # Test suite
│   ├── unit/            # Unit tests
│   ├── integration/     # Integration tests
│   └── e2e/             # End-to-end tests
├── docs/                 # Documentation
├── examples/             # Example code
├── tools/                # Demo and simulation tools
├── alembic/             # Database migrations
├── templates/           # Jinja2 templates
└── config/              # Configuration files
    └── nginx/           # Nginx configuration files

Requirements

• Python 3.12+
• Docker and Docker Compose (for easy deployment)
• PostgreSQL (Docker Compose handles this automatically)
• Google OAuth credentials (for Admin UI)
• Gemini API key (for AI features)

Contributing

We welcome contributions! Please see our Development Guide for:

• Setting up your development environment
• Running tests
• Code style guidelines
• Creating pull requests

Important: Database Access Patterns

When contributing, please follow our standardized database patterns:

# ✅ CORRECT - Use context manager
from database_session import get_db_session
with get_db_session() as session:
    # Your database operations
    session.commit()

# ❌ WRONG - Manual management
conn = get_db_connection()
# operations
conn.close()  # Prone to leaks

See Contributing Guide for details.

Admin Features

Multi-Tenant User Access

Users can belong to multiple tenants with the same email address (like GitHub, Slack, etc.):

• Sign up for multiple publisher accounts with one Google login
• Different roles per tenant (admin in one, viewer in another)
• No "email already exists" errors - users are tenant-scoped

Migration: Database schema updated with composite unique constraint (tenant_id, email). See alembic/versions/aff9ca8baa9c_allow_users_multi_tenant_access.py

Tenant Deactivation (Soft Delete)

Deactivate test or unused tenants without losing data:

How to deactivate:

1. Go to Settings → Danger Zone
2. Type tenant name exactly to confirm
3. Click "Deactivate Sales Agent"

What happens:

• ✅ All data preserved (media buys, creatives, principals)
• ❌ Hidden from login and tenant selection
• ❌ API access blocked
• ℹ️ Can be reactivated by super admin

Reactivation (super admin only):

POST /admin/tenant/{tenant_id}/reactivate

Self-Signup

New users can self-provision tenants:

• Google OAuth authentication
• GAM-only for self-signup (other adapters via support)
• Auto-creates tenant, user, and default principal
• Available at /signup on main domain

Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Documentation: docs/

License

Apache 2.0 License - see LICENSE file for details.

Related Projects

• AdCP Specification - Protocol specification
• MCP SDK - Model Context Protocol tools
• FastMCP - MCP server framework