v0.1.0

v0.1.0

First MVP release, bringing the research workflow into one application.

• Data setup: Upload and check CSV files, choose a prediction target, and optionally fill missing predictor values using training data only.
• Research workflow: Choose agent skills for hypothesis generation, data preparation, machine learning, SHAP/LIME interpretation and expert explanation. Use research documents or LLM knowledge only. Create, name and review Hypothesis, Data and ML skills with AI, with automatic numbering for duplicate names; Data and ML skills retain their selected split and algorithm.
• Model settings: Start without LLM credentials and configure Azure Foundry in the GUI, or use .env defaults. Enter the Azure deployment name and endpoint directly. Checked GUI settings become the default for new experiments and AI skill creation; a complete model configuration is required before execution.
• Reproducible runs: Configure train/test splitting, pipeline order and replicates. Save execution seeds and apply each run's seed to both the split and model random state.
• Execution controls: Follow live agent activity and logs, pause/resume supported executions, and retry interrupted work while preserving saved progress and results.
• Results and usage: Browse completed outputs by pipeline and replicate. Results show saved Interpretation output, Conclusions show saved Explanation output, and Usage shows Motoro-reported usage and cost estimates. Subtle New badges highlight unread updates in all three views.
• Usability and reliability: Simplify experiment setup and responsive layouts, with AI skill settings opening directly from Create with AI; improve input validation, saved-state recovery and automated regression checks.
• Installation: Start the application with Docker, Motoro 0.13.0 and the included Heart Failure example. Align package and Docker release versions and exclude local design photos from source packages.

Updated MVP snapshot

This v0.1.0 release has been refreshed to commit 6ee424e43bea9484b13dba5c5f82acb3800d7a31. Both attached source archives were regenerated from this exact commit.

• Automatic selection of an available web port.
• Simplified model setup and Explanation skill choices.
• Custom Interpretation and Explanation skills through the library, upload/write and AI creation.
• Configurable Explanation report length (default 250 words), with bounded revisions and recorded length checks. The length remains approximate.
• Correct back-navigation arrows for Results and Conclusions.
• Updated installation and release links.