1. Pick a Model
Model ID
Type any HuggingFace model ID (e.g., microsoft/Phi-3-mini-4k-instruct)
Connect Model
Enter a model ID and click Connect.
2. Test It
🔄 LOREIN Pipeline (Plan → Generate → Verify)
🧭 Model Router
🔄 LOREIN Pipeline (Plan → Generate → Verify)
🧭 Model Router
Ask something
Run Pipeline
Clear
Final Response
Pipeline output will appear here...
How the Pipeline Works
MIRROR Thinking
— Decomposes your query into Goals/Reasoning/Memory threads
Model Generates
— The model answers with the MIRROR context as background
Truth Check
— Scans for hedging, contradictions, uncertainty
Governor
— Safety screen (mutual advancement principle)
Memory + Journal
— Logs everything to the hash-chained ledger