MCP setup guides for every client

The config reference for each major MCP client — where the config file lives, the exact format, supported transports, and the common failures — plus how to route every tool call through PolicyLayer. See the Quick start for the end-to-end gateway setup.

Claude Code

.mcp.json

CLI-based coding assistant by Anthropic

Cursor

.cursor/mcp.json

AI-native code editor, VS Code fork

Claude Desktop

claude_desktop_config.json

Desktop app for conversational AI

Windsurf

~/.codeium/windsurf/mcp_config.json

AI development environment (Cascade)

VS Code

.vscode/mcp.json

GitHub Copilot with MCP support

Cline

cline_mcp_settings.json

Autonomous coding agent for VS Code

OpenAI Codex

~/.codex/config.toml

OpenAI's coding agent CLI

Gemini CLI

~/.gemini/settings.json

Google's AI CLI with MCP support

Zed

settings.json → context_servers

High-performance editor with AI agent

Continue

~/.continue/config.yaml

Open-source AI code assistant

Take your agents live. Without losing control.

Route your MCP traffic through PolicyLayer. Every tool call is checked against your policy before it runs: allow, deny, or require approval. Per-identity grants. Full audit log. Live in minutes.

TAKE YOUR AGENTS LIVE →

Instant setup, no code required.

46,500+ MCP servers and 515,000+ tools scanned and risk-classified.