| # Roadmap |
|
|
| ## Phase 0: Repository Bootstrap |
|
|
| Status: complete. |
|
|
| - Initialize Git repository. |
| - Connect GitHub remote. |
| - Add README and contributor docs. |
| - Add stack-neutral project structure. |
| - Add lightweight repository health check. |
|
|
| ## Phase 1: Source Intake |
|
|
| Status: complete. |
|
|
| - Add Python validation engine. |
| - Wire MiniCPM through an OpenAI-compatible endpoint. |
| - Add CLI harness with seeded test ideas. |
| - Run the spiral and judge whether pressure is sharp or generic. |
|
|
| ## Phase 2: Core Implementation |
|
|
| Status: in progress. |
|
|
| - Build the Gradio app after the engine clears the validation gate. |
| - Embed the custom Iris canvas/cards UI through HTML/CSS/JS. |
| - Keep the four-pressure demo flow smooth and screen-recordable. |
| - Preserve local MiniCPM as the load-bearing AI path. |
| - Preserve the "pressure, not answer" product rule. |
|
|
| ## Phase 3: Release and Hosting |
|
|
| Status: pending. |
|
|
| - Deploy the Gradio app as a Hugging Face Space under the hackathon |
| organization. |
| - Add deployment configuration. |
| - Add CI checks. |
| - Add demo video and social post links to the README. |
| - Prepare release notes, field notes, and operational docs. |
| - Optional: investigate llama.cpp serving for an extra badge only if it does not |
| risk the working local MiniCPM path. |
|
|
