External AI agents for Pega application development
Pega Platform™ supports application development from external AI-assisted development environments using Model Context Protocol (MCP). By connecting external enterprise AI agents to Pega authoring services, teams can use natural-language interactions to understand, document, test, extend, and modify applications. This capability also provides flexibility for organizations that have standardized enterprise coding assistants and supported development environments while continuing to use the model-driven development approach of Pega Platform.
External AI agents and Pega AI Assistant use the same underlying authoring capabilities but offer different development experiences. Pega AI Assistant provides an embedded experience in Pega Infinity Studio™. External agents operate from compatible development environments such as Claude Code, GitHub Copilot, OpenCode, Codex, and Devin AI.
External-agent support enables developers to continue using approved tooling while accessing Pega authoring capabilities. This approach is useful for teams that:
- Use enterprise AI coding assistants
- Want to use existing AI subscriptions and model providers
- Need AI-assisted development for Pega Platform applications
- Prefer development environments outside Infinity Studio
Pega Infinity Authoring plugin architecture
The Pega Infinity Authoring plugin connects compatible external development environments to Pega applications. You can use the plugin with GitHub Copilot, Claude Code, and other compatible agentic development tools.
The plugin contains two primary components:
- Pega MCP server exposes Pega authoring capabilities as MCP tools that external agents can invoke.
- Pega Infinity skills supply Pega-specific knowledge, including authoring recipes, Rule Type schemas, library references, and development best practices.
The MCP tools perform authoring activities such as creating and updating Rules, running Rules, generating test cases, retrieving Case Type information, and searching application assets. The plugin uses Pega Platform APIs and services to perform these operations within the authorization context of the user.
The following figure shows the primary components of the Pega Infinity Authoring plugin and their interactions:
Installing and configuring the authoring plugin
Before you can use an external AI agent to work with a Pega application, install the Pega Infinity Authoring Plugin and configure a connection to the target application. Complete the following high-level steps to install the plugin and connect it to a Pega application:
- Confirm that the plugin supports the target Pega Platform release. The plugins work with Pega Platform 26 without additional configuration. For information about required hotfixes for earlier releases, see the plugin repository documentation mentioned at the end of this section.
- Select a compatible development environment. Confirm that the selected environment supports the Pega Infinity Authoring plugin.
- Download or clone the Pega Infinity AI plugins repository from GitHub.
- Follow the README file and the instructions for the selected plugin to install the plugin and Pega Infinity Skills.
- Run the pega-setup skill. The skill checks the connection to Pega Infinity™ and guides you through initial configuration and troubleshooting.
- Configure the connection and authentication details for the target Pega application as described in the repository documentation.
- Verify the connection by asking the agent to inspect the application context before you request an authoring change.
- Create or modify a small artifact in a Branch and then review the proposed change before you use the plugin for a larger project.
For detailed installation and configuration instructions, see Pega Infinity Authoring plugins.
Governance and best practices
External-agent development follows the same governance principles as AI-assisted authoring in Infinity Studio. All authoring occurs in Branch workspaces, and you must explicitly approve changes before incorporating them into the application. This process maintains traceability while you use AI-assisted development.
As a Lead System Architect (LSA), consider reviewing AI-generated changes for:
- Alignment with the application architecture
- Opportunities to reuse existing assets
- Security requirements
- Compliance with development standards
- Test coverage and application behavior
When you work with external agents, apply the following best practices:
- Consider external agents when your organization has standardized on enterprise coding assistants. With external-agent development, teams can continue using approved coding assistants and AI models while accessing Pega authoring capabilities through the Pega Infinity Authoring plugin.
- Verify platform compatibility before installation. Confirm that the target Pega Platform release supports the Pega Infinity Authoring plugin and any required prerequisites before enabling external-agent development.
- Plan for model consumption and licensing costs. When you use third-party coding agents, your organization's licenses and subscriptions govern token consumption. Benefits associated with Pega-managed language models do not apply to external coding assistants.
- Extend your agent's capabilities by building custom skills through the Pega Infinity plugins framework, which also includes a growing set of Pega-authored skills.
For more information, see Developing applications with external agents.
Check your knowledge with the following interaction:
This Topic is available in the following Module:
If you are having problems with your training, please review the Pega Academy Support FAQs.
Want to help us improve this content?