llms.txt
# Metomic
> Metomic is the security control point between AI and enterprise data: MCP Gateway for agent-to-tool traffic, inference hooks for model traffic, Shadow AI discovery for the tools nobody approved. Every exchange is inspected, then allowed, redacted, or blocked, and always logged.
Key facts:
- Metomic is a single security control point between AI and enterprise data. Five capabilities sit beneath it: the MCP Gateway, inference hooks, Shadow AI discovery, AI data-in-motion inspection, and policy enforcement.
- MCP Gateway: secures and governs agent-to-tool and agent-to-data traffic. It sits between AI agents and the tools and models they call, inspecting every Model Context Protocol (MCP) request in real time and enforcing least-privilege access.
- Inference hooks: inspect and control data before it reaches an AI model. They are interception points wired into the model call path, covering AI traffic that never passes through an MCP server, such as direct LLM API calls, embedded copilots, and custom agent frameworks.
- Shadow AI discovery: finds and governs unmanaged AI usage. Shadow AI is the use of AI agents, MCP servers, copilots, and browser AI tools that were never approved by security. Metomic surfaces this layer so teams can govern it rather than block it.
- AI data-in-motion inspection: detects sensitive data in prompts, model responses, and agent exchanges. Metomic reads the content of each exchange, not just its metadata, so a request that looks fine by sender and destination is still caught when it carries something it should not.
- Policy enforcement: every exchange is allowed, blocked, redacted, or held for human approval in real time, and audited. Metomic keeps a full audit trail of every agent action, streamed to your SIEM.
- Deployment is hosted and preconfigured, with no agents on endpoints.
- Metomic is SOC 2 Type II certified (AICPA) and rated 4.8 on G2.
## Product
- [Platform](/platform/): Full platform overview, from data discovery to real-time enforcement
- [MCP Gateway](/solution/mcp-gateway/): Inspect every agent-to-tool MCP request in real time and allow, redact, hold, or block it
- [Claude Inference Hooks](/solution/claude-inference-hooks/): Agentless AI security for Claude Enterprise. Metomic is the AI security server behind Anthropic's Claude Inference Hooks, returning real-time DLP verdicts on Claude Enterprise prompts and tool calls, covering AI traffic that never passes through MCP. Nothing is installed on endpoints and no network or proxy changes are required; connect the webhook and verdicts start appearing on live traffic in under a minute, fully live in under 30 minutes
- [Shadow AI](/solution/shadow-ai/): Get visibility into the AI tools and agents nobody approved, and control what they touch
- [Data Loss Prevention](/solution/data-loss-prevention/): Content-aware DLP across SaaS and AI
- [Data Security Posture Management](/solution/data-security-posture-management/): Find and fix sensitive-data exposure at rest
- [All Integrations](/integration/all-integrations/): Salesforce, Slack, Jira, Google Drive, ChatGPT, and more
## Resources
- [What Is an MCP Gateway?](/resource-centre/what-is-an-mcp-gateway/): Definition, core capabilities, and how it differs from an API Gateway, DLP, or CASB
- [What Is Agent Governance?](/resource-centre/what-is-agent-governance/): The runtime controls that decide what an AI agent can access, do, and prove
- [AI Agent Governance: The Company Harness](/resource-centre/ai-agent-governance-company-harness/): Metomic founder Ben van Enckevort on governing a company whose actors outnumber its people, and on the company harness that holds the rules
- [What Is AI Data in Motion?](/resource-centre/what-is-ai-data-in-motion/): Why protecting data in transit through AI requests differs from data-at-rest tooling
- [MCP Gateway vs. API Gateway](/resource-centre/mcp-gateway-vs-api-gateway/): Why the two solve different problems and most teams need both
- [MCP Gateway vs. Inference Hooks](/resource-centre/mcp-gateway-vs-inference-hooks/): Two deployment paths for inspecting AI data in motion, their tradeoffs, and why most teams need both
- [AI Gateway vs. Secure Web Gateway](/resource-centre/ai-gateway-vs-secure-web-gateway/): Why an SWG's browser-traffic visibility doesn't cover agent-to-tool traffic over MCP
- [Agent Security vs. AI Security](/resource-centre/agent-security-vs-ai-security/): Why securing the model doesn't automatically secure what it's connected to
- [MCP Security: A CISO's Guide to Agent-Tool Risk](/resource-centre/mcp-security-ciso-guide/): What MCP is, where it breaks, and how to secure it
- [Shadow AI: What It Is and How to Control It](/resource-centre/shadow-ai/): Definition, risks, and a governance playbook
- [Resource Centre](/resource-centre/): Guides, case studies, whitepapers, and reports
- [Blog](/resource-category/blog/): Research and practical guidance on AI and data security
## Company
- [About Metomic](/about/): Company background and team
- [Security Centre](/security-center/): Certifications, subprocessors, and security practices
- [Book a demo](/book-a-demo/): See the MCP Gateway on your own stack
## Optional
- [Careers](https://apply.workable.com/metomic/): Open roles
- [Partners](/partners/): Partner with Metomic
- [Metomic on AWS Marketplace](/metomic-on-aws-marketplace/): Procure through AWS
