Sylvia started as a personal ML playground for learning, but it's evolving into a mono-repo of independent ML microservice apps. Each app is self-contained but can share common libraries, infrastructure, and tooling.
Sylvia can be launched using CLI, GUI, or an API backend (future integration). Currently, a stub backend is available for safe testing.
# Launch CLI with the stub backend
python main.py --mode cli --backend stubCLI Features:
-
/switch [profile]— switch active personality profile -
/hybrid [profile:weight,...]— set weighted hybrid personalities -
+/-— provide feedback for last response -
exit— quit the CLI -
Save conversations interactively after each message
# Launch GUI with the stub backend
python main.py --mode gui --backend stubGUI Features:
-
Send messages to Sylvia via a chat box
-
Switch personality profiles or set hybrid weights
-
Debug logs and response times displayed in real-time
-
Stubbed visualization panel (Matplotlib) showing placeholder data
-
Uvicorn-powered API for remote interaction:
python main.py --mode api -
Provides REST endpoints for sending messages and switching profiles
-
Full model integration coming in future updates
-
🧩 Modular Apps: Each app is isolated and can evolve independently.
-
🔄 Shared Utilities: Avoid code duplication and promote reusability.
-
🚀 Scalable: Add new ML apps easily while maintaining clean structure.
-
📊 Experiment-Friendly: Notebooks, data, and models are organized per app for reproducibility.
-
💬 Interactive Interfaces: CLI and GUI allow real-time interaction with SylviaBot.
-
🧪 Safe Testing: Stubbed backend returns canned responses and prevents runtime errors from the full Personality engine or plotting issues.
-
Re-integrate the full Personality engine and advanced model backends.
-
Enable Matplotlib-based visualization in GUI with live personality updates.
-
Expand CLI/API backends for local and remote models.
-
Add more ML apps and microservices while maintaining modularity.
Sylvia is moving from a single experimental playground to a full ecosystem of ML tools — modular, scalable, interactive, and experiment-ready.