For the complete documentation index, see llms.txt. This page is also available as Markdown.

Agentic Development

BoxLang is built for AI agents and developers working together. Learn how to set up your project, skills, MCP servers, and IDE for agentic development.

BoxLang is a software productivity platform built for developers and the AI agents working beside them. Agents build better software on platforms that are predictable, well documented, and complete. This section shows how to set up a BoxLang project so your agent is productive from the first prompt.

🎯 Why BoxLang Works Well for Agents

Trait
What it means for an agent

Batteries included

Files, HTTP, JDBC, async, caching, scheduling, templating, logging, and AI ship with the platform. The agent wires up fewer libraries and makes fewer choices.

Conventions

Application.bx lifecycle, module layouts, configuration in boxlang.json, and consistent BIF naming give agents patterns to follow.

Skills

Installable SKILL.md playbooks from skills.boxlang.io teach the agent how to perform specific BoxLang tasks the idiomatic way.

Docs over MCP

The documentation is available to agents as MCP servers, so they read current docs instead of guessing.

Live introspection

With BoxLang+, the bx-mcp module lets an agent inspect a running server through MCP, with access control.

Java underneath

100% Java interoperability gives the agent the whole JVM ecosystem when it needs it.

Fast validation

boxlang check validates syntax without running code, so agents can verify every edit.

One runtime, many targets

The same code runs on the CLI, web servers, containers, and serverless platforms.

🧰 The Agentic Toolkit

AI SetupSkills & GuidelinesValidate Your CodeBoxLang MCPBuild an App End to EndBoxLang IDEAgent-Ready Checklist

🚀 Fastest Path

  1. Install BoxLang.

  2. Install the BoxLang developer skills: npx skills add ortus-boxlang/skills/boxlang-developer. See Skills & Guidelines.

  3. Add a project AGENTS.md and connect the docs MCP servers. See AI Setup.

  4. Have your agent validate every edit with boxlang check. See Validate Your Code.

  5. Building a web application? Use ColdBox and its coldbox ai install wizard.

  6. Want your agent to see a running server? Try BoxLang MCP.

  7. Want your agent to use the app it builds, in a real browser? See Build an App End to End with bx-playwright.

🤖 Building AI Features, Not Just With AI

Agentic development is about using agents to write your application. BoxLang also lets you build AI into your application with chat, agents, tools, memory, RAG, and MCP through the BoxLang AI module.

Last updated

Was this helpful?