Every tool call logged. Your rules on every call.
Add ContextGate to Claude Code, Cursor, VS Code or Codex as one MCP server. Every tool call through it is logged, checked against your policies, and blocked, warned on or redacted when it should be. Free.
Then sign in to ContextGate when your app asks (in Claude Code, run /mcp). New to ContextGate? Signing in creates your free account. Setup in ContextGate walks you through Codex, Gemini CLI, Claude and ChatGPT too.
The installer signs you in through your browser, finds Claude Code, Claude Desktop, Cursor and VS Code on your machine, and adds ContextGate to the ones you pick.
Each tool call is checked against your workspace policies, then logged with its arguments, the redacted result and what your policies decided.
Switch off individual tools on a connection. Switched-off tools are hidden from your AI app and refused if called.
Add apps such as Gmail, GitHub, Salesforce and HubSpot from the app catalog, or any remote MCP server. Switch off the tools you don't want used: they're hidden from your AI apps and refused if called.
It takes minutes, and you keep working in the AI app you already use.
One command in Claude Code, or one click in Cursor or VS Code. Sign in when your app asks; that creates your free account if you don't have one.
Add apps such as GitHub, Slack and Gmail from the app catalog, or any remote MCP server, including your own. Tool lists sync from each one.
Switch off tools you don't want used. Add policies: plain-English governance checks, PII redaction, secrets blocking and blocked patterns.
Keep using your AI app. Every tool call through ContextGate is logged and checked, then allowed, warned on, redacted or blocked.
Any other MCP client: add a remote (HTTP) MCP server with the URL https://api.contextgate.ai/mcp and sign in. A client that can't sign in can use an API key instead; Setup in ContextGate shows how.
Checks run on a tool's arguments before it runs, and on its results before your AI app sees them. Each one blocks, warns or logs. Policies apply to every tool call in your workspace. Start from one of 16 templates or write your own.
Personal data in tool results is replaced with placeholders before your AI app sees it. When the app sends a placeholder back, the tool gets the real value. 40 built-in types, plus your own patterns.
A tool call whose arguments contain an API key, AWS key, JWT or other secret is blocked, and the secret is kept out of the log. Add your own regex and keyword patterns.
Write a rule in plain English. A model that ContextGate manages judges each call against it, so you don't need model keys. Scope a check to specific tools.
See which AI app called which tool, for which user, with what arguments, and what your policies decided. Results are stored redacted. Filter by app, connection, tool, status, policy, user and date. Attention lists the last 7 days of blocked and warned calls in one place.
{ "query": "Acme Ltd" }
{ "name": "Jo Smith", "company": "Acme Ltd", "email": "[EMAIL_ADDRESS_REDACTED_3f9a1c2e]", "note": "Refund to [UK_SORT_CODE_REDACTED_7c1e04ab] [ACCOUNT_NUMBER_REDACTED_b52d9e10]" }
ContextGate sees the tool calls that go through it. It doesn't see your AI app's built-in tools, such as Claude Code's own file edits and terminal commands, or your chats.
Building an agent in-house? Give it an API key and it connects to ContextGate like any other MCP client. Its tool calls are checked and logged the same way as your editor's.
Create an API key in your workspace for an agent or script. Revoke it when you no longer need it.
Point the agent at the remote MCP URL from the Setup page and authenticate with the key.
Your workspace policies check every tool call it makes, just as they do for your editor.
Its tool calls appear in Activity with their arguments, redacted results and policy outcomes, next to everyone else's.
Invite your teammates. Their AI apps connect to the same workspace, under the same policies, and every tool call lands in the same Activity log.
Add teammates as Owner, Admin, Member or Viewer. Each person signs in to ContextGate from their own AI apps.
The Agents page lists the AI apps and API keys using ContextGate, who signed them in, and when each was first and last seen.
Tool calls, blocks and warnings, calls per day by outcome, top tools and calls by AI app.
The last 7 days of blocked and warned calls in one list.
| Role | |
|---|---|
| you@company.com | OWNER |
| sarah@company.com | MEMBER |
| contractor@ext.io | VIEWER |
Free. No card required. Add ContextGate to Claude Code, or sign up in your browser.