User feedback collection on canvas interaction and AI model performance.

## Share a real PenEcho canvas

We would love to see what people try with PenEcho. A small example can reveal more than a long feature request and may inspire the next recognition, interaction, or model-support improvement.

Reply with any of the following:

- a screenshot or short recording of the canvas;
- the question, equation, diagram, or idea you explored;
- the model and executor you used, such as API or Codex CLI;
- what worked well;
- what felt confusing, slow, or limited;
- what you wanted PenEcho to understand or draw next.

Examples from learning, mathematics, physics, software architecture, research, and everyday visual thinking are all welcome.

Before posting, remove API keys, private endpoints, personal notes, names, or other sensitive information. By sharing an example, you help us understand which real workflows deserve the most attention.