IDE Integrations with MCP¶
Use CloudAEye through your IDE's coding agent to describe changes and review uncommitted changes or open pull requests. The shared plugin provides skills, scripts, and an OAuth-protected MCP connection.
Choose your IDE¶
| Client | Setup guide |
|---|---|
| VS Code | VS Code |
| Cursor | Cursor |
| PyCharm | PyCharm |
| IntelliJ IDEA | IntelliJ IDEA |
| Codex CLI or desktop | Codex |
| Claude Code CLI | Claude Code Plugin Setup & User Guide |
Prerequisites¶
- A Git repository and a CloudAEye account with Code Review access.
- An activated coding agent with available usage allowance or API credits.
- Git, Python 3, curl, and Bash accessible to the agent.
Install CloudAEye from the plugin repository using your client's setup guide. The complete package includes the supporting scripts required by its skills.
Endpoint and authentication¶
| Environment | MCP URL |
|---|---|
| Hosted | https://api.cloudaeye.com/mcp |
Configure your client to connect to https://api.cloudaeye.com/mcp.
First workflow¶
- Open the repository you want to review.
- Load the skills and authenticate MCP using your client guide.
- Select init and complete repository setup if requested.
- Select inspect, security, or review.
For PR review, select review and add a number such as #412.
Limitations¶
- Agent access can add subscription or API costs alongside CloudAEye.
- Exhausted agent quotas can block new workflows even when MCP is healthy.
- Agent orchestration adds latency and variability compared with direct API calls.