Topview Motion StudioGet a launch video for about $3, not thousands
Promoted
Launching today
VLM Camera — Chrome 機能拡張
Turn your PC into a 100% secure local AI camera. Free,WebGPU
2 followers
Turn your PC into a 100% secure local AI camera. Free,WebGPU
2 followers
Turn your PC into a secure, private AI camera with free on-device Chrome extensions. We release 2 tools: VLM Camera for local VLM (LFM, Florence-2, SmolVLM) analysis & crop-OCR, and Face Recognition Manager for shared-PC utilization logging. Both run 100% locally in your browser with zero cloud uploads or history tracking. Optional encrypted backups go directly to your Google Drive, flexible camera controls, mobile notifications, and rich parameter customization.
Hello Product Hunt community! 🚀
I'm thrilled to introduce VLM Camera and Face Recognition Manager—two Chrome extensions designed to give you powerful AI vision capabilities right inside your browser, with zero cloud dependencies and absolute privacy
.
■ Why we built this: Most AI applications today harvest your data. If you use a camera or upload photos, those files sit on someone else’s server. We believe AI should serve you, not exploit your privacy. By running everything locally, these extensions process your camera feed entirely on your machine
. Nothing is saved as history, and nothing is uploaded to any third party
. Even your backups go directly to your personal Google Drive with optional encryption
.
■ What you can do:
VLM Camera: Use vision-language models (LFM, SmolVLM or Florence-2) to capture, describe, and analyze scenes, or run a highly efficient "Detect -> Crop -> OCR" pipeline (perfect for reading text or license plates in real-time)
.
Face Recognition Manager: Track shared-PC utilization via lightweight face recognition
. It doesn’t save face images or video—only encrypted, local feature vectors to help teams coordinate
. It runs flawlessly and fast on CPU-only machines
.
■ Key Technical Highlights:
VLM Speed: Runs local inference in just 6 seconds using the LFM 450M model on setups like RTX 2060 or Ryzen 5 5625U
.
CPU Friendly: While heavy VLMs prefer WebGPU, lightweight tasks like face recognition and crop OCR run on standard CPU setups with no GPU required
.
15 Languages Natively Supported: Accessible to teams worldwide with fully translated interfaces
.
Mobile Coming Soon: A smartphone web app version is releasing shortly!
Both tools are completely free to use locally. Cloud option is a paid feature (less than $1 per device. This spends server costs for managing the authentication for integration. No data is transferred or stored by our server).
I would love for you to try them out and share your thoughts! What models or pipelines should we add next?
Happy private AI.