Give eyes to yourAI agents

Visor gives LLMs verified visibility into live iOS and Android apps so they can check real implementations against Figma, catch pixel drift, and iterate until the UI matches.

Get Started
visor run checkout-smoke.json --platform ios --app-id com.example.app
Live Runtime
Appium-backed iOS and Android targeting for real implementation checks.
Captured Evidence
> Run Artifacts
• screenshots/app-opened.png
• sources/app-opened.xml
• summary.json
> Reports
timeline.log
junit.xml
report.html
Verification Loop
// use Visor to verify what the LLM shipped
1. Implement the screen
2. Capture live evidence
3. Check against Figma
4. Iterate until drift is gone

Everything agents need to verify mobile UI

Evidence, control, and repeatability for Figma-aware implementation loops

Live App Evidence

Capture screenshots and UI source from running apps so agents reason from real runtime state instead of assumptions.

Cross-Platform Control

Drive iOS and Android through Appium with a single CLI for taps, navigation, typing, and runtime targeting.

Figma Verification

Use Visor as the verification layer in your design loop to catch pixel drift and confirm the implementation matches Figma.

Repeatable Runs

Validate scenarios, collect artifacts and reports, evaluate assertions, and benchmark determinism across repeated executions.

Simple, powerful CLI

Concrete commands for evidence capture, scenario runs, and repeatable verification

$ visor validate checkout-smoke.json
{
"status": "ok",
"next_action": "run",
"data": {
"valid": true,
"issues": []
}
}

Scenario parsed and ready for execution

How it works

From live runtime to repeatable Figma verification in four steps

01

Install and target your app

Install the TypeScript CLI, point Visor at Appium, and choose the iOS or Android runtime your agent needs to verify.

02

Capture real evidence

Record screenshots, UI source, and run artifacts from the live app so the agent works from actual implementation state.

03

Verify against Figma

Use Visor's evidence inside your LLM workflow to spot drift from the intended design and decide what the implementation must change.

04

Repeat until aligned

Rerun scenarios, review reports, and keep the loop hands-off until the shipped UI matches the design with confidence.

Ready to verify what your agent built?

Use Visor to capture evidence from live mobile apps, check alignment against Figma, and keep the implementation loop repeatable.

Open source
MIT License
TypeScript CLI