Integrations

How Certiv fits your environment

Certiv is not a catalogue of connectors. It runs where the agents run, so it deploys through the device management you already have, sees whatever your engineers install, and enforces policy against any model behind it.

In short

Certiv deploys through existing MDM (Jamf, Intune, Kandji, Workspace ONE) onto macOS, Windows and Linux. Because Scout observes agent activity at the endpoint, coverage does not depend on a connector existing per tool. It sees any coding agent, browser agent or MCP server, and the Policy Engine enforces against hosted and locally-run models alike.

01 · Deploy

Through the MDM you already run

Scout installs in minutes through existing device management. No kernel extensions, though it does require a single reboot. From there it keeps a light footprint (minimal CPU, unnoticeable latency) and fails open by default so it never becomes the reason engineering slows down.

  • Jamf
  • Microsoft Intune
  • Kandji
  • Workspace ONE

Runs on macOS, Windows and Linux workstations.

02 · Observe

Whatever agent your engineers reach for

Coverage is not per-tool. Scout captures agent activity at the endpoint, so a new agent does not need a new connector before it becomes visible, including the ones nobody filed a ticket for.

  • Claude Code
  • Codex
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Gemini CLI
  • Amazon Q
  • Zed
  • Browser agents
  • MCP servers

Regardless of which platform or framework the agent runs on.

03 · Enforce

Against any model, hosted or local

The Policy Engine evaluates the actions an agent takes rather than the model behind it. That is what lets enforcement cover hosted models, local models and shadow agents equally, including models that never generate network traffic at all.

  • OpenAI
  • Anthropic
  • Google
  • Meta Llama
  • Qwen
  • DeepSeek
  • Ollama and other local runtimes

A locally-run model is governed by exactly the same policies.

Why there is no connector list to maintain

Every control that works by inspecting traffic inherits the same limit: it governs what it can see, and it can only see what routes through it. That is why those tools need an integration per vendor, and why a new agent is invisible until one is built.

Certiv sits in the execution path on the endpoint instead. Coverage never depends on traffic crossing a proxy, or on us having heard of the tool your engineer installed this morning.

Straight Answers

What Teams Ask About Fitting Certiv In

Expand to view common questions.

Do you need an integration for every agent?
No, and that is the point of running on the endpoint. Scout observes AI agent activity regardless of which platform or framework those agents run on, so coverage does not wait on a connector being built. A tool that shipped last week is visible the day someone installs it.
How does Certiv get deployed?
Through the device management you already run. Scout deploys in minutes, requires no kernel extensions (though it does need a single reboot), keeps CPU usage minimal with unnoticeable latency, and fails open by default so productivity continues uninterrupted.
Does Certiv replace our gateway, CASB or EDR?
No. They compose. A gateway remains the control point for API key custody and model routing; EDR still watches the device. Certiv covers what those structurally cannot: local models, shadow agents, off-network users, and the endpoint context behind every tool call. Certiv’s decision records also give a gateway’s traffic logs the missing context of which agent acted, and why.
What about models running entirely offline?
They are covered identically. Enforcement happens in the execution path on the endpoint rather than on the wire, so an agent using a locally-run open-weight model is evaluated against the same policies as one calling a hosted frontier model.

See Certiv on Your Own Endpoints

Deploy in minutes. See every agent, then control what happens next.

Book a Demo

Keep reading

Scout, the endpoint agent

What actually gets installed, how light it is, and what it does and does not look at.

Local & open-weight models

The clearest case for endpoint-native enforcement: models that never touch the network.

Certiv vs. AI gateways

How the two layers compose, and what a gateway structurally cannot reach.