Reproduce reported issues
Give the agent the same app, device context, screenshots, touches, and session history your team would normally reconstruct by hand.
Let agents test and verify their own work
The bottleneck isn't writing the code. It's the human who has to manually test it.
App
2026-08-05 02:42 · 1m 20s · 41,161 logs
Analyse Share Export
Telemetry
TIMELEVELMESSAGE
No annotations Annotations pinned during capture appear here with their timeline range.
A real captured session. Pick another from the list, scrub the timeline, or open a tab.
Live when the issue is happening. Replayable after it is gone.
Maintenance rarely happens in one sitting. Ansight keeps the runtime evidence useful across the full handoff: from the person who finds the issue, to the agent that investigates it, to the teammate who reviews the result.
Agent tool calls
ansight_get_app_state foreground ansight_take_screenshot 480×1044 · 28 KB ansight_get_logs 41,161 lines ansight_get_telemetry 3 channels Ansight SDK 1.1.0-preview.1
A local path from your agent to your app.
Keep production boundaries intact. Ansight is designed for trusted development and QA workflows, with SDK capabilities and tool access controlled by the app.
AI agent Studio Ansight SDK Your app
Read the setup guide - 01
Install Studio
Run the local desktop bridge that connects your app, your team, and your coding agent.
- 02
Add the SDK
Choose the package for your stack and keep its maintenance tools behind your developer-build guard.
- 03
Pair once
Scan the one-use QR. Later development sessions reconnect automatically on your trusted network.
- 04
Connect your agent
Add Ansight as a local MCP server for Codex, Claude, Cursor, or another compatible client.
Give your agent the runtime context maintenance requires.
Ansight does not replace your coding agent. It gives that agent a reliable way to observe the app, act through approved tools, and verify the evidence around its work.
Inspect the live runtime
Read UI trees, focused controls, app state, preferences, files, databases, and SDK-exposed tools from the running app.
Correlate the evidence
Line up logs, screenshots, telemetry, navigation, touches, and app-owned artifacts on one maintenance timeline.
Exercise real workflows
Let an agent invoke guarded app tools, push test state, inspect the result, and keep the investigation moving without a screen-share handoff.
Verify the outcome
Capture the post-change UI and runtime evidence so the agent can check what actually happened instead of assuming the edit worked.
Preserve the handoff
Share captured sessions, annotations, evidence, and analysis with the next developer, tester, or agent—without rebuilding context.
One maintenance workflow across the apps you ship.
The SDK details vary by platform. The agent-facing workflow stays familiar: connect, inspect, correlate, act through approved tools, and verify.
Keep building when you are away from your development machine.
Start with Ansight Studio, pair a development build, and give your coding agent the live runtime context it needs. Continue from your phone or tablet when the companion app arrives.