Browse categories
Explore
Fiverr Pro
English
$
USD
I will build an autonomous android bot with ai screen vision
Sharif Rafid
About this gig
️ AUTONOMOUS ANDROID BOTS THAT ACTUALLY SEE THE SCREEN
Standard automation breaks the moment an app updates its layout. Mine doesn't.
Most Android bots read UI "nodes" invisible labels behind buttons. The
second a developer ships a redesign, or the app draws on a canvas (games,
banking apps, protected enterprise tools), those bots go blind.
I build bots that LOOK at the screen. By combining native Android
Accessibility Services with applied computer vision (OpenCV + optional
LLM vision), the bot identifies elements visually, understands context,
and taps the right spot even after a UI change.
WHAT YOU GET
- AI screen-vision detection (resilient to UI updates)
- Native Accessibility Service integration
- Dynamic coordinate mapping (no hardcoded taps)
- Editable config change targets without touching code
- Clean source code + documentation
- Retry logic & error handling so it doesn't crash silently
PERFECT FOR
- QA & regression testing teams
- Data collection / scraping workflows
- Repetitive in-app task automation
- Internal enterprise tooling
Get to know Sharif Rafid
Sharif Rafid
Full Stack Web And Mobile Apps Developer
- FromBangladesh
- Member sinceFeb 2017
- Avg. response time1 hour
- Last delivery1 month
Languages
Bengali, English
Founder & Senior Android Engineer | 8+ Years Exp.
Since 2017 I have built and secured Android apps for clients worldwide, and I now run a small software studio. I focus on the device-level work most developers avoid: kiosk and lock-task deployments, accessibility-service automation, BLE and IoT integrations, and OWASP MASVS security audits where every finding maps to a control your auditor will recognise. I also build the Next.js and Node.js backends that sit behind those apps. Message me before ordering and I will tell you honestly whether your project is a fit.
My Portfolio
FAQ
Will the bot break when the app updates?
That's the entire point of the vision approach. Node-based bots break on redesigns; my Standard/Premium tiers detect elements visually and adapt.
Does it need root access?
No. It runs on Accessibility Services, which work on non-rooted devices.
Can it work on games or canvas-based apps?
Yes — that's where vision shines, since those apps have no readable UI nodes. Premium tier recommended.
Do I get the source code?
Yes, on Standard and Premium. Basic is delivered as a ready-to-run build.
Can you deploy it to run 24/7?
Yes, via the Cloud/VPS deployment extra (emulator or device farm).
Is this legal?
Automation itself is legal. I only build for legitimate use cases — testing, internal tooling, and authorized data collection. I decline anything that breaches ToS or law.
