Connect any MCP-compatible AI to your infrastructure.
OneIQ registers as a standard MCP connection, so whichever AI tool your team has already standardized on can query a governed, read-only context cloud instead of raw infrastructure.
01 · Microsoft AI Ecosystem
Microsoft AI ecosystem
Many of our customers and partners run on Microsoft for AI, whether that's Copilot licenses sitting unused, an internal Foundry deployment, or Copilot inside VS Code. OneIQ registers as a standard MCP connection across all of it, so the capacity and tooling you've already paid for finally has something real to query.
-
Microsoft Foundry
If your team builds and hosts its own agents in Microsoft Foundry, point them at OneIQ over MCP. They inherit governed, read-only infrastructure context cloud, with no custom connector to build and nothing to justify to your security team.
-
VS Code / GitHub Copilot
GitHub Copilot's MCP support means an engineer can ask an infrastructure question directly inside VS Code, such as which environment a service is actually deployed to and whether it's healthy, and get a governed, structured response instead of switching tools or paging someone on-call.
-
Copilot Studio / Microsoft 365 Copilot
This is the flagship integration: the Copilot your organization already trusts and has already paid for, now able to answer infrastructure questions in Teams or Copilot Chat because OneIQ sits behind it as a standard MCP connection. See the full walkthrough on our Copilot for IT Teams and Copilot for Channel Partners pages.
02 · Public Hosted AI
Public hosted AI
You've already chosen a frontier model for everything else your team does, and there's no reason infrastructure questions should require a different one.
-
Claude
Connect Claude to OneIQ over MCP and it can reason over governed infrastructure context the same way it reasons over any other tool call: read-only, scoped to the Agent Zone you define, and fully audit-logged.
-
ChatGPT
ChatGPT can query a governed Agent Zone via MCP without OneIQ ever exposing production credentials, raw telemetry, or anything outside the boundary you've defined. The model can provide an answer without opening a direct connection to your live environment.
03 · Private / Partner-Hosted AI
Private and partner-hosted AI
Many organizations, and most managed-service partners, keep their AI model in-house rather than trusting a third-party API. OneIQ works wherever the model runs, as long as it speaks MCP. Self-hosted and partner-hosted AI query OneIQ's governed, read-only context cloud without touching live production systems.
-
Open WebUI
Open WebUI is the self-hosted front end many MSPs and IT teams already run in front of local or private models. Register OneIQ as an MCP tool inside that instance, and every model you serve through it can query a governed Agent Zone instead of a raw infrastructure API.
04 · VS Code AI Extensions
VS Code AI extensions
Plenty of teams get their AI assistance through a VS Code extension rather than a dedicated IDE or standalone client. Each extension below reaches OneIQ through an MCP connection, so the experience is consistent no matter which one your team has standardized on.
-
Cline
Add OneIQ as an MCP server in Cline and it gets scoped, read-only infrastructure context inside the exact VS Code setup your team already uses. There's no new tool to learn and no separate credentials to manage.
-
Roo Code
Connect Roo Code to OneIQ over MCP so an infrastructure question returns a governed, structured response instead of a raw data dump the model then has to interpret on its own.
-
Continue
Register OneIQ as an MCP tool in Continue for infrastructure-grounded answers that sit right alongside the codebase context it's already using.
05 · Local CLI / Agent
Local CLI and agent tools
Increasingly, the first place someone asks an infrastructure question isn't a dashboard, it's the terminal they're already working in. Every CLI-based agent below connects through a standard MCP client, allowing them to query infrastructure through OneIQ's governed read-only context cloud.
-
Claude Code
An engineer using Claude Code to work on a service can ask what infrastructure that service actually runs on, including capacity, health, and upstream dependencies, and get a governed answer through the same MCP connection without leaving the terminal or opening a second tool.
-
Codex
Give Codex a governed, read-only view of your environment over MCP so any infrastructure-adjacent answer it gives, whether that's sizing a deployment, checking a dependency, or flagging a risk, is grounded in real, current state instead of a guess based on what's in the repo.
-
Gemini CLI
Gemini CLI connects to OneIQ through the same MCP interface as every other client here, so a team standardized on Gemini gets identical scoped, governed infrastructure context with no separate integration path required.
06 · AI IDE / Agent
AI IDEs and coding agents
AI-native IDEs have their own built-in agents that reach well beyond autocomplete, and those agents increasingly need to know things about the infrastructure the code they're writing will actually run on. Registering OneIQ as an MCP server gives them that context without ever handing them real infrastructure credentials.
-
Cursor
Register OneIQ as an MCP server in Cursor so its agent can ground capacity, cost, and dependency-risk answers in what you actually run instead of inferring from code alone, all inside the same editor your team already lives in.
-
Devin Desktop
Devin Desktop's agent can query a governed Agent Zone over the same MCP connection, so infrastructure-aware answers show up inside your existing workflow rather than requiring a separate lookup somewhere else.
07 · Private Desktop AI
Private desktop AI
For teams and partners who want a model running entirely on local hardware, with no external API calls, OneIQ still applies the same governance rules over the same MCP interface. Local doesn't mean unscoped.
-
LM Studio
Run a local model in LM Studio and connect it to OneIQ over MCP the same way you would any hosted client. The model stays local and gets governed, read-only infrastructure context from OneIQ.
Any MCP client, one governed layer
Don't see your
AI tool listed?
If it supports the Model Context Protocol, it can connect to OneIQ. Tell us what you're running and we'll walk you through it.