This guide takes you from a clean workstation to a first paired session.

If you want an AI agent to install and integrate Ansight into a .NET app, please use this skill /skills/dotnet/ansight-install-dotnet.md.

Copyable prompt:

Please use this skill https://ansight.ai/skills/dotnet/ansight-install-dotnet.md to install and integrate Ansight into my .NET application. Set up the base Ansight package, add any required tool packages, initialize Ansight with host pairing, keep the integration scoped to local development, and tell me what manual Studio or pairing steps remain.

Prerequisites

  • A trusted desktop machine for running Ansight Studio.
  • A development build of your app on the same local network as Studio.
  • A .NET mobile app if you want to follow the SDK steps in this docs set.
  • A local AI client that can talk to MCP if you want agent-driven inspection.

1. Install Studio

Install Ansight Studio on your desktop and launch it before you start wiring the app.

Studio is the local control plane for:

  • issuing pairing configs
  • exporting QR pairing payloads
  • capturing and browsing sessions
  • exposing the local MCP bridge to agents

Use a developer-controlled machine, not a shared workstation.

2. Confirm the Local MCP Bridge

Studio exposes MCP locally at:

http://127.0.0.1:45125/mcp/

Studio also listens on:

http://localhost:45125/mcp/

Open Studio and confirm the MCP panel shows the bridge as running before you connect an agent.

3. Point Your Agent at Studio

Every MCP client has its own config shape, but the value you need is the Studio endpoint above.

Use the local URL in your agent’s MCP config:

{
  "ansight": {
    "url": "http://127.0.0.1:45125/mcp/"
  }
}

Notes:

  • Use the exact config format your MCP client expects.
  • Studio only accepts requests from localhost.
  • The client handles MCP session initialization; you only provide the endpoint.

For client-specific setup steps, read MCP Configuration.

4. Install the SDK Packages

Install the base .NET package:

dotnet add package Ansight

Add tool-suite packages only when you need them, and keep them scoped to local development builds.

Examples:

dotnet add package Ansight.Tools.VisualTree
dotnet add package Ansight.Tools.Database

5. Initialize Ansight in the App

Start with a minimal runtime setup:

using Ansight;

var options = Options.CreateBuilder()
    .WithFramesPerSecond()
    .WithHostPairing()
    .Build();

Runtime.InitializeAndActivate(options);

From here you can add custom channels, JPEG session capture, tool suites, or custom pairing loaders as needed.

6. Issue a Pairing Config in Studio

In Studio:

  1. Open the apps workspace.
  2. Register the app name and app identifier.
  3. Issue a pairing config for that app.
  4. Choose a duration that matches the workflow you need.

Studio stores the signed config locally and can also turn it into a QR payload for device-side pairing.

7. Pair the App

There are two common flows:

  • Developer pairing: bundle ansight.json, enable the developer pairing target, and let the app auto-connect to the preferred profile.
  • QR pairing: scan the Studio QR code in the app, then hand the scanned payload to Runtime.HostPairing.ConnectFromPayloadAsync(...).

The dedicated pairing guide is here:

8. Start Capturing Sessions

Once the app is paired and connected:

  • Studio shows live and saved sessions.
  • The dashboard exposes logs, telemetry, screenshots, and device metadata.
  • Agents connected over MCP can inspect Studio-owned data and, when enabled, live app tools.

Next Steps

  • Read Studio for the desktop workflow.
  • Read MCP Configuration for client-specific setup in Codex, Cursor, and Claude Code.
  • Read SDK Overview for the current SDK lineup.
  • Go straight to .NET SDK if you are integrating a .NET app.
  • Read Security before you enable any remote tools.