Over Shoulder

Over Shoulder turns expertise into instant AR training. Pros perform a task once; AI captures actions, voice, and context to create guides. Rookies learn by doing with real-time coaching.

Comment

The education prompt clicked with us, as showing is the best way to teach. We started building different visual guides, and understood that the teaching is hard..

So, being lazy, we asked: what if we just demonstrate a skill once, and the system builds the whole contextual guide? - That's how we found our direction, and a simple question became the core of Over Shoulder - The platform that turns expertise into guidance.

In OverShoulder, Pros are not recording tutorials; they simply perform wearing SPECS, and the system extracts expertise and stores it in our Skills Library (https://overshoulder.familybusiness.studio/).
As Skill appears in our Library on the web, it can be reviewed and edited if needed.

We remove teaching from teaching.

Rookies can learn anything by doing: getting help when they are stuck, seeing guides in context, or just scrap the plan and just browse and ask the guide questions.

Not all skills use static objects; ML models allow us to attach contextual labels to moving objects like rabbits, cards and finger boards, allowing you to look at the skill you are learning.

For now, OverShoulder transfers experience from human to human, but in future, it can share skills from human to robot, transfer online videos and books into visual guides, to be integrated into the school programs and health industry. We believe that in future, anyone can do anything with OverShoulder

Built With

  • cloudflare-d1-sqlite-gemini
  • cloudflare-pages
  • css
  • depth-cache/camera-depth-apis
  • html
  • javascript
  • lens-studio
  • openai-realtime-asr
  • plain-typescript
  • python
  • remote-service-gateway
  • rest
  • snap-spectacles
  • spectacles-interaction-kit
  • spectacles-uikit
  • sql
  • vite
  • ypescript
Share this project:

Updates