← All experiments
POC // 009 Specialist agents

StickyShip

Photograph a whiteboard. Review the extracted notes, then turn the approved drafts into real GitHub Issues.

Photographed sticky-note whiteboard used for the StickyShip backlog experimentAPPLIED AI
THE IDEA

A closer look.

Turn a simple three-column sticky-note board into a project backlog. OpenCV.js finds coloured notes and Tesseract.js reads each crop in the browser. Optional local Ollama helps structure the extracted text.

Review and correct every draft before approving a specific GitHub destination. The resulting issue links stay beside the original board.

How it works

  1. 01

    Upload a clear photo of a three-column board.

  2. 02

    Review detected notes and correct their OCR text and status.

  3. 03

    Optionally structure the drafts with local Ollama.

  4. 04

    Approve the exact repository and create the reviewed GitHub Issues, or export JSON.