Spaces:
Running
Running
| # Soap Lab β Agent Trace | |
| ## Example: Agentic Recipe Optimiser | |
| **Input:** "300g coconut oil, 400g olive oil β optimise this recipe for me" | |
| **Step 1 β Parse:** coconut oil: 300g, olive oil: 400g | |
| **Step 2 β Score:** | |
| - Cleansing: 27.9 (HIGH β target 12-22) | |
| - INS: 103.5 (LOW β target 136-165) | |
| **Step 3 β Iteration 1:** | |
| - Cleansing too high β transfer 35g coconut β olive | |
| - Recipe: coconut 265g, olive 435g | |
| **Step 4 β Rescore:** | |
| - Cleansing: 24.6 (still high, improved) | |
| - INS: 97.6 (still low) | |
| **Step 5 β AI Explanation:** | |
| - Model: Qwen/Qwen2.5-3B-Instruct via Featherless AI | |
| - Explains changes in plain English to user | |
| ## Architecture | |
| - Chemistry engine: deterministic Python (no AI) | |
| - AI layer: explanation only, never calculates | |
| - Model: Qwen2.5-3B-Instruct (Tiny Titan eligible) | |