Delphi & AI Agents

Your ERP already holds the knowledge of the business. Now it can talk to AI agents.

Delphi in the Age of Agents

Build MCP servers and connect systems and ERPs to Artificial Intelligence

Build the bridge between your Delphi systems and AI agents with MCP, controlled capabilities, and human approval.

Notify me when it is available

This book is currently available in Brazilian Portuguese.

EditionIn preparation2026Book languagePortugueseBrazilFormatsComing soonComing soon

About the book

A complete view of the book.

Delphi in the Age of Agents shows how to bring corporate systems and ERPs closer to artificial intelligence agents without granting unrestricted access to their internal complexity. The book introduces agents, tools, context, and the Model Context Protocol before guiding readers through the architecture and implementation of an MCP server written in Delphi.

The journey uses a fictional MiniERP to turn real rules and operations into explicit capabilities: queries may run automatically, preparation tasks create drafts, and critical actions require human approval. By progressing through tools, resources, prompts, contracts, authorization, testing, and observability, readers learn to build a controlled bridge between business knowledge and AI—from the first server to the path required for a dependable solution.

The challenge

Does this situation sound familiar?

Delphi systems and ERPs concentrate essential rules, data, and operations, but handing that knowledge directly to an agent can expose complexity, expand permissions, and turn a promising integration into an operational risk.

The book’s approach

A path from intention to practice.

Build an MCP server in Delphi, understand agents, tools, and context, and connect a MiniERP to controlled queries and operations with explicit contracts, human authorization, observability, and limits proportional to risk.

“Expose capabilities, not complexity. Keep control over every decision.”

By the end of the book, you will be able to:

✓Understand how agents use tools, context, and the MCP protocol

✓Design the architecture of an MCP server written in Delphi

✓Implement tools, resources, and prompts with explicit contracts

✓Connect agents to real ERP queries and business capabilities

✓Control critical operations with authorization and action limits

✓Prepare tests, observability, and evolution beyond the prototype

Questions that guide the reading

Answers begin with the right questions.

Questions explored through context, criteria, and examples.

How does an agent discover and use capabilities from a Delphi system?

How can an MCP server be built without an external library?

How can ERP operations be exposed without offering generic SQL access?

Which controls separate an interesting demo from a dependable integration?

Inside the book

A structured journey that helps you move forward.

Concepts and applications presented in a coherent sequence.

Agents and MCP

Understand the paradigm shift, the role of context, and how MCP connects models to systems.

MCP server in Delphi

Model and implement the architecture, transport, and message cycle of an MCP server in Delphi.

Tools, resources, and prompts

Expose tools, resources, and prompts as clear, validated, and discoverable business capabilities.

MiniERP integration

Apply the server to a MiniERP, offering intelligent queries without opening generic SQL access.

Security and authorization

Classify operations by risk and incorporate authorization, human approval, and explicit limits.

Testing and operations

Create evidence with logs, tests, and observability to evolve from an experiment to a dependable solution.

Content in preparation

Planned book structure

Explore the 4 parts and 12 planned chapters. Titles may be refined during editorial review.

View chapters12 chapters

Part I — The paradigm shift

Delphi in the age of agents
How agents, tools, and context work
MCP: the protocol connecting AI to systems

Part II — Building the MCP server in Delphi

Architecture of an MCP server with Delphi
The first MCP server
Tools, resources, prompts, and the Delphi client

Part III — From the server to the corporate system

Meet the book's MiniERP
Intelligent queries and business context
Operations controlled by agents

Part IV — Safe operation and evolution

Security, authorization, and action limits
Observability, testing, and reliability
From prototype to an agent-oriented MiniERP

What makes this book different

Depth to understand. Clarity to keep moving forward.

An MCP server built in Delphi to reveal the protocol and architecture

A continuous project with a fictional MiniERP and business capabilities

Operations classified as reading, preparation, or critical action

Security, human approval, and observability integrated throughout the journey

Who this book is for

Knowledge that becomes action.

Delphi developers, architects, and technical leaders who maintain or evolve corporate systems and ERPs will find a direct approach connected to concrete problems.

✓ Delphi developers maintaining corporate systems and ERPs✓ Architects who need to integrate applications with AI agents✓ Technical leaders responsible for security and governance✓ Professionals who want to understand and build MCP servers

Before you begin

What you need to benefit from the book

✓Intermediate experience with Delphi and Object Pascal

✓Familiarity with APIs, JSON, and business systems is helpful, but MCP is introduced from the ground up

An informed choice

This may not be the right book if…

—Readers seeking only ready-made prompts or a generic introduction to AI

—Readers planning to give agents unrestricted database access

Why this author, on this subject

Experience that gives the reading context.

An Embarcadero MVP and Delphi developer for more than 20 years, Régys connects his experience in corporate systems, ERPs, and software architecture to secure integration between applications and artificial intelligence agents.

Upcoming release

Receive resources and updates about Delphi in the Age of Agents.

Upcoming release