Live Stream Buyer Detection App

I want a single mobile application that runs smoothly on both iOS and Android and quietly watches TikTok LIVE, Instagram Live, and Facebook Live while those apps are open on the user’s phone. The moment a viewer types a name, username, or a comment that signals buying intent—such as “price?” or “I want one”—the app should pick it up through on-device OCR, flag it in real time, and log it for later follow-up.

Because the video never leaves the device, the OCR engine must read the text directly from the screen overlay, using whichever accessibility or screen-capture approach each platform allows, without disrupting the broadcast. Accuracy and speed are crucial; a delay of a few seconds could mean a lost sale.

Deliverables
• Universal iOS/Android build (native or a single codebase such as Flutter or React Native)
• Integrated OCR module tuned for live-stream text
• Real-time matching rules for names, usernames, and intent phrases (extensible via settings)
• Lightweight dashboard that shows detected leads as they appear and exports them to CSV or Google Sheets
• Clear setup instructions and source code

I’ll test by running live sessions on all three platforms and checking that interested buyers are captured and listed correctly.

---
來源：Freelancer.com | 出價次數：14 | 剩餘時間：6 日 | 專案類型：固定預算