---
license: mit
title: api
sdk: docker
emoji: 🚀
colorFrom: blue
colorTo: red
short_description: d
---
# 🧠 Savant RRF Φ12.5 — Meta-Logic & Resonance API

Savant RRF Φ12.5 is a production-ready FastAPI system for **evaluating, ranking, and improving AI-generated reasoning** using the Resonance of Reality Framework (RRF).

It combines **embedding geometry, spectral analysis, and meta-logic classification** into a single inference pipeline.

---

## 🔗 Live API

https://antonypamo-apisavant3.hf.space
---

## ⚡ Core Capabilities

- 🧠 **Meta-Logic Evaluation**  
  Scores reasoning quality beyond surface semantics

- 🔍 **Semantic Reranking**  
  High-performance batched similarity ranking

- 🧩 **Φ-node Ontology Mapping**  
  Deterministic cognitive-state classification (8 nodes)

- 🔁 **Closed-loop Improvement (Φ12.5)**  
  Evaluate → refine → optimize responses

---

## 📦 Models

| Component | Model |
|----------|------|
| Embedder | `antonypamo/RRFSAVANTMADE` |
| Meta-Logic | `RRFSavantMetaLogicV2` |
| Feature Dim | 15 |
| Runtime | CPU (GPU optional) |

---

## 🧩 Φ-Node Ontology

The system maps inputs into one of 8 deterministic cognitive states:

| Index | Node |
|------|------|
| 0 | Φ0_seed |
| 1 | Φ1_geometric |
| 2 | Φ2_gauge_dirac |
| 3 | Φ3_log_gravity |
| 4 | Φ4_resonance |
| 5 | Φ5_memory_symbiosis |
| 6 | Φ6_alignment |
| 7 | Φ7_meta_agi |

Mapping is derived from **spectral coherence, energy distribution, and phase dynamics**.

---

## 🚀 API Endpoints

### `GET /`
Service discovery

---

### `POST /evaluate`
Evaluates reasoning quality

**Request**
```json
{
  "prompt": "Explain what a smoke test is",
  "answer": "A smoke test is a minimal validation..."
}