# How to Use AI Agents in IntelliJ IDEA With ACP

ACP, developed through JetBrains and Zed collaboration, simplifies AI agent integration into IDEs.

- **Source:** [The JetBrains Blog](https://brifing.si/source/782eebd1-7be8-49cc-b27c-e552516691b2)
- **Published:** 2026-08-12
- **Topic:** [Data & AI](https://brifing.si/c/data-ai)
- **Original article:** <https://blog.jetbrains.com/idea/2026/08/how-to-use-ai-agents-in-intellij-idea-with-acp/>
- **On Brifing:** https://brifing.si/s/56bdbead-11b5-47ee-a10f-0ed0ddd8c907

## What is ACP?

The Agent Client Protocol (ACP) establishes a standard communication layer between IDEs like IntelliJ IDEA and AI coding agents. Developed jointly by JetBrains and Zed, ACP creates a unified integration framework for coding agents, similar to how Language Server Protocol (LSP) standardizes language support. This eliminates the need for custom plugins per agent. ACP operates over JSON-RPC, normalizing communication between the IDE and agents while maintaining flexibility.

## Built-in Agents

IntelliJ IDEA ships with three ACP-compatible agents ready for immediate use: Codex, Claude Agent, and Junie. These pre-installed agents require no additional setup. Each agent maintains unique workflow characteristics—some support planning modes while others feature slash commands or specialized authentication flows. ACP preserves these differences while providing consistent interaction management within the IDE environment.

## Extending via ACP Registry

IntelliJ IDEA's ACP Registry offers additional agent options accessible through Settings > Tools > AI Assistant > Agents. The registry provides all necessary metadata for agent installation, updates, and removal. Since each agent maintains independent licensing, service terms, and privacy policies, reviewing these details before granting repository access is crucial. This centralized approach simplifies agent management without requiring plugin maintenance.

## Custom Agent Integration

Internally developed ACP-compatible agents can be registered via the ~/.jetbrains/acp.json file. This configuration specifies the agent's executable path and required arguments for ACP mode activation. IntelliJ IDEA's 'Add Custom Agent' action automates file creation and editing. Custom agents typically target specific organizational needs rather than general distribution, making this lightweight integration method particularly valuable for enterprise environments.

## Technical Architecture

ACP standardizes the IDE-agent interaction while allowing each agent to maintain its own models, behaviors, and tools. For local agents, IntelliJ IDEA spawns subprocesses communicating via JSON-RPC over standard I/O streams. During operation, prompts flow to agents while progress updates, file operations, and permission requests return to the IDE. This bidirectional communication occurs within ACP's standardized protocol boundaries.

## Model Context Protocol (MCP)

Complementing ACP, the Model Context Protocol (MCP) delivers additional context and tools to agents. MCP enables access to user-configured servers and the integrated IntelliJ MCP server, among other resources. By providing richer contextual awareness, MCP improves agent decision-making accuracy for coding tasks. Agent settings allow configuration of which MCP servers to utilize, offering granular control over available resources.

## Multi-Agent Workflows

Teams can employ multiple agents specialized for different tasks—one might excel at code refactoring while another performs better at complex algorithm generation. ACP facilitates seamless integration of diverse agents into IntelliJ IDEA, giving developers workflow flexibility. When switching between agents, project management and code review processes remain uninterrupted within the IDE, maintaining productivity across specialized tools.

## ACP's Strategic Value

ACP is emerging as a critical standard accelerating AI-assisted coding adoption. By eliminating custom plugin development overhead, it enables teams to evaluate and integrate best-of-breed agents rapidly. The protocol's clear boundary between IDE and agent responsibilities supports independent evolution of both components. This approach fosters growth and diversification within the AI coding assistant ecosystem while maintaining developer workflow consistency.

## Sources covering this story

- https://blog.jetbrains.com/idea/2026/08/how-to-use-ai-agents-in-intellij-idea-with-acp/ — 

---

_Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._