This project was created entirely through vibecoding — a development approach where all code, documentation, tests, and configuration are written by AI (LLM-based agents and assistants), under human direction and review.
- 100% AI-generated: Every line of code, every test case, every documentation file, and every configuration was produced by AI tooling (e.g., ChatGPT, Claude, OpenCode agents).
- Human-in-the-loop: A human provided the initial requirements (see
prompt.txt), reviewed outputs, and directed the AI through iterative refinement — but never wrote a single line of production code by hand. - No manual coding: No traditional "hand-written" code exists in this repository. The human acted as product owner, reviewer, and QA — not as an implementer.
- Initial prompt → MVP: The project went from a natural-language specification (
prompt.txt) to a working MVP (backend capture server, image processing, WebSocket streaming, frontend dashboard) entirely through AI-driven iteration. - Iterations: Multiple rounds of AI refinement were used to fix bugs, improve blur quality, add access control, and polish the frontend — all without manual code edits.
All files in this repository are covered by this declaration, including but not limited to:
backend/— Python server, screenshot capture, image blurring, WebSocket handling, testsfrontend/— HTML, CSS, JavaScript dashboardREADME.md— Project documentationprompt.txt— Original requirements specificationVIBECODING.md— This file- Configuration, test fixtures, and build artifacts
MIT — same as the rest of the project. See README.md for details.