How CC GUI Turns IDEA into a Full‑Featured Claude Code & Codex Workbench (Free, Open‑Source)

Problem: Switching between IDEA and AI terminals

IDEA users edit code in the IDE while interacting with Claude Code or Codex in a separate terminal, requiring frequent context switches that interrupt workflow.

Solution: CC GUI plugin

CC GUI (formerly Claude Code GUI) is an MIT‑licensed JetBrains plugin that embeds Claude Code and Codex CLI agents into a dedicated IDEA side panel, providing a visual workbench for AI‑assisted coding.

Core capabilities

Toggle between Claude Code and Codex without reinstalling.

Use the @file command to attach files, images, or describe requirements, supplying richer context than raw CLI input.

Display changes in IDEA’s Diff panel with clickable links to the exact line.

History search, favorites, export, and token‑cost statistics show token usage and equivalent USD cost.

Built‑in slash commands and MCP server extensions enable custom workflows.

Comparison with official ACP integration

JetBrains supports Claude Code via the Agent Client Protocol (ACP), which runs lightweight agents inside the IDE. CC GUI takes a different route: it provides a full visual workbench with session management, image input, Skill system, and MCP support, targeting users who prefer an all‑in‑one GUI rather than a minimal CLI bridge.

Installation (three steps)

Open Settings → Plugins, search for “CC GUI”, and install the version with the highest download count (≈385.6k downloads).

Configure a provider using one of four methods:

Create an Anthropic API key.

Import an existing ~/.claude/settings.json file.

Use the community tool cc-switch for one‑click provider selection.

Specify a custom third‑party endpoint.

Start a conversation from the right‑hand panel, e.g., “Analyze the current project structure”.

Slash command /plan : AI as architect first

In /plan mode the AI scans the project, reads relevant classes (e.g., PaymentService), extracts key points from integration docs, and returns a detailed design document that lists files to modify, new classes, required SDKs, signature locations, and callback handling. /plan 设计下银联的接入方案 The generated plan includes:

Files to modify.

New classes to add.

Dependencies (SDKs) required.

Where signatures and encryption should be performed.

How callbacks and asynchronous notifications are handled.

After review, switch back to /agent mode to apply changes.

Token consumption and cost visibility

Simple 5‑turn dialogue: ~15 K tokens.

Code review of ~100 lines: ~25 K tokens.

Medium‑size project planning with /plan: ~40 K tokens.

The plugin shows token count and equivalent USD cost in a dedicated statistics panel.

Additional AI‑powered features

AI Commit : Generates Conventional Commits messages from staged diffs, up to five times faster than manual typing.

AI Review : Scans a diff before a commit or PR, flagging boundary conditions, null‑pointer risks, potential infinite loops, and naming issues.

MCP Server : Configure address, headers, and authentication once; the same mcp.json works for Claude Desktop and Cursor. A Chrome DevTools MCP extension is recommended for front‑end debugging.

Comparison with Qoder (Alibaba) and CodeBuddy (Tencent)

Nature : CC GUI is a GUI shell for CLI agents; Qoder and CodeBuddy are independent AI agents.

Open‑source : CC GUI is MIT‑licensed; the others are closed‑source.

Model support : CC GUI works with Claude, Codex, and custom models; Qoder and CodeBuddy ship with built‑in models.

Entry barrier : CC GUI requires an existing Claude/Codex subscription; Qoder offers a free quota after registration; CodeBuddy provides a free personal tier.

Strengths : CC GUI reuses existing subscriptions and benefits from community contributions; Qoder is deeply optimized for the Java ecosystem; CodeBuddy integrates tightly with the Tencent cloud stack.

All three can be installed simultaneously without conflict.

Limitations

Output quality depends on the relevance and correctness of the supplied context (e.g., proper use of @file).

AI lacks knowledge of business constraints; clear project boundaries must be defined by the user.

Generated code must be reviewed; the plugin shortens the path but cannot replace human judgment.

Recommended starter tasks: adding unit tests, generating boilerplate code, outlining module logic, or converting screenshots into initial code.

Resources

GitHub repository: https://github.com/zhukunpenglinyutong/jetbrains-cc-gui

JetBrains Marketplace page: https://plugins.jetbrains.com/plugin/29342-cc-gui-claude-or-codex-