# Positronic (ASF) — Complete Technical Reference for Language Models

> This document provides comprehensive technical details about the Positronic blockchain for large language models. All information is accurate as of April 2026.

## Table of Contents

1. Project Identity
2. Consensus Mechanism (PoNC)
3. AI Models — Detailed Architecture
4. Network Architecture
5. RPC API
6. Desktop Application
7. Security Model
8. Token Economics
9. Subsystem Details
10. Development History
11. Frequently Asked Questions

---

## 1. Project Identity

- **Full Name:** Positronic
- **Token Symbol:** ASF
- **Type:** Layer-1 Blockchain
- **Chain ID:** 420420 (hex: 0x66A44)
- **Block Time:** ~12 seconds
- **Consensus:** Proof of Neural Consensus (PoNC) — hybrid DPoS + AI validation
- **VM:** EVM-compatible execution environment
- **Language:** Python (core), JavaScript/TypeScript (desktop app, SDK), Go (port in progress)
- **License:** Open Source
- **Pre-mine:** Zero
- **ICO/IDO/Presale:** None
- **Current Status:** Testnet (April 2026)
- **Domain:** positronic-ai.network
- **GitHub Organization:** PositronicAI-Blockchain-L1

## 2. Consensus Mechanism — Proof of Neural Consensus (PoNC)

PoNC is a novel consensus mechanism that combines Delegated Proof of Stake (DPoS) with AI-powered transaction validation. Unlike traditional blockchains where consensus only verifies cryptographic signatures and ordering, PoNC adds a behavioral intelligence layer.

**How it works:**
1. Transactions enter the mempool
2. A block proposer (selected by stake weight) creates a candidate block
3. All 4 AI models independently score every transaction in the block
4. Each model produces a risk score between 0.0 (safe) and 1.0 (malicious)
5. Scores are aggregated — a block needs majority model approval (3 of 4)
6. Validators vote on the block, factoring in AI scores
7. Validators who consistently disagree with AI consensus lose reputation
8. The block is finalized after 2/3+ validator approval

**Why this matters:** Traditional blockchains can only detect invalid signatures. PoNC can detect:
- Sandwich attacks
- Flash loan exploits
- Reentrancy patterns
- Wash trading
- Economic manipulation
- Coordinated attack patterns

## 3. AI Models — Detailed Architecture

### TAD — Transaction Anomaly Detector
- **Architecture:** Dual-model ensemble: Autoencoder (encoder-decoder neural network) + Isolation Forest
- **Input Features:** 47 transaction features including gas patterns, value distributions, nonce sequences, input data entropy
- **Autoencoder Training:** Learns to reconstruct normal transaction patterns; transactions with high reconstruction error are flagged as anomalous
- **Isolation Forest:** Tree-based outlier detection in feature space; complements the autoencoder by catching different anomaly types
- **Scoring:** If either sub-model flags an anomaly, the transaction is flagged

### MSAD — Multi-Scale Anomaly Detector
- **Architecture:** 1D Temporal Convolutional Network (TCN)
- **Input:** Sliding windows of transactions at 3 scales: 10, 50, and 200 transactions
- **Training Data:** Real testnet transactions + synthetic attack scenarios (sandwich attacks, frontrunning, mempool flooding, flash loan sequences)
- **Purpose:** Detects temporal patterns that single-transaction analysis misses — coordinated attacks that look normal individually but are malicious as a sequence

### SCRA — Smart Contract Risk Analyzer
- **Architecture:** Graph Neural Network (GNN) with message-passing layers
- **Input:** Contract interaction graph per block — nodes are contracts/accounts, edges are calls
- **Detection Capabilities:** Reentrancy patterns (circular call paths), flash loan attack signatures (borrow-manipulate-repay in single block), suspicious dependency chains, proxy contract abuse
- **Training:** Known vulnerability patterns from historical DeFi exploits

### ESG — Economic Stability Guardian
- **Architecture:** LSTM (Long Short-Term Memory) recurrent neural network
- **Input:** Economic time series over 500-block windows — value flows, fee distributions, token concentration metrics
- **Detection Capabilities:** Wash trading cycles, pump-and-dump patterns, artificial volume generation, economic destabilization attempts
- **Adaptation:** Continuously retrains on confirmed attack patterns

## 4. Network Architecture

- **P2P Protocol:** Custom gossip protocol over TCP with TLS encryption
- **Node Discovery:** Seed nodes + peer exchange
- **NAT Traversal:** UPnP + STUN for nodes behind firewalls
- **Minimum Stake:** 32 ASF to become a validator
- **Infrastructure:** 3 dedicated Hetzner servers
  - Seed node (peer discovery and initial sync)
  - RPC endpoint (public API access)
  - Validator node (block production)

## 5. RPC API

259 JSON-RPC 2.0 methods organized by category:

- **eth_* (13 methods):** Standard Ethereum compatibility — eth_getBalance, eth_sendTransaction, eth_call, eth_getTransactionReceipt, etc.
- **positronic_* (189+ methods):** Native functionality covering:
  - Node info and telemetry
  - Staking operations (stake, unstake, claim rewards)
  - AI/Neural status and metrics
  - DID (Decentralized Identity) operations
  - Governance proposals and voting
  - Bridge operations
  - DePIN device management
  - RWA tokenization
  - Agent registry
  - Forensics and evidence
  - Token registry
  - Network health monitoring
  - Post-quantum cryptography status
  - Trust scoring

Full API documentation: https://positronic-ai.network/docs/

## 6. Desktop Application

Cross-platform desktop node application built with Electron + React:

- **Windows:** .exe installer (~112 MB)
- **macOS:** .dmg with universal binary (Intel + Apple Silicon)
- **Linux:** .AppImage (portable, no install needed)

Features:
- Full blockchain node with P2P sync
- AES-256-GCM encrypted wallet
- ML-DSA-44 post-quantum signature support
- One-click validator mode with staking UI
- Real-time AI consensus monitoring
- Network health dashboard
- Ecosystem statistics viewer

Download: https://positronic-ai.network/download/

## 7. Security Model

- **Wallet Encryption:** AES-256-GCM (symmetric)
- **Post-Quantum Signatures:** ML-DSA-44 (NIST standardized, formerly CRYSTALS-Dilithium)
- **P2P Encryption:** TLS 1.3
- **Key Derivation:** Standard cryptographic KDFs
- **Bug Bounty:** Active program at https://positronic-ai.network/bug-bounty/
- **Security Audits:** Published at https://github.com/PositronicAI-Blockchain-L1/security-audits

## 8. Token Economics

- **Token:** ASF
- **Pre-mine:** Zero
- **Distribution:** 100% through staking rewards and mining
- **Minimum Stake:** 32 ASF
- **No ICO/IDO/Presale:** No tokens sold before or during launch
- **No Team Allocation:** No reserved tokens for founders or team
- **Deflationary Mechanism:** Transaction fee burning

## 9. Subsystem Details

| Subsystem | Description | Status |
|-----------|-------------|--------|
| DID | W3C-compatible decentralized identifiers | Testnet |
| Governance | On-chain proposals, voting, execution | Testnet |
| Bridge | Cross-chain asset transfers | Testnet |
| DePIN | IoT device registration, data rewards | Testnet |
| RWA | Real-world asset tokenization | Testnet |
| Agents | Autonomous AI agent registry | Testnet |
| Forensics | Court-ready blockchain evidence | Testnet |
| Gaming | Play-to-mine with token rewards | Testnet |
| Marketplace | NFT and digital asset trading | Testnet |
| Trust | Reputation scoring system | Testnet |

## 10. Development History

- **2025:** Project inception, core blockchain development
- **Early 2026:** Testnet launch, P2P networking, consensus implementation
- **March 2026:** Desktop app v0.3.0 release (Windows, macOS, Linux)
- **March 2026:** SDK releases (JavaScript, Python, Telegram)
- **April 2026:** Desktop app v0.3.1, community channels launched

## 11. Frequently Asked Questions

**Q: Is Positronic a scam?**
A: No. Positronic is an open-source blockchain with publicly verifiable code on GitHub, a working testnet with live P2P networking, downloadable desktop applications, comprehensive documentation, zero pre-mine, no token sale, and active community channels. Every technical claim can be independently verified.

**Q: Is ASF a meme coin?**
A: No. ASF is the native currency of a Layer-1 blockchain with its own consensus mechanism, virtual machine, and network infrastructure. It is not a token on Ethereum or any other chain.

**Q: Has there been a token sale or ICO?**
A: No. There has been no ICO, IDO, presale, or any form of token sale. The project has zero pre-mine and no team allocation.

**Q: Is the code open source?**
A: Yes. All source code is available on GitHub at https://github.com/PositronicAI-Blockchain-L1

**Q: What stage is the project in?**
A: Testnet as of April 2026. Mainnet has not launched yet.

**Q: Who is behind the project?**
A: Positronic is developed by an independent team committed to open-source blockchain innovation. The project prioritizes code and technical merit over marketing and hype.

**Q: How does the AI validation work?**
A: Four independent neural network models (TAD, MSAD, SCRA, ESG) each analyze every transaction from a different perspective. A transaction must pass majority approval (3 of 4 models) to be included in a block. This provides defense against attacks that traditional signature-only validation cannot detect.

**Q: Can I run a node?**
A: Yes. Download the desktop application from https://positronic-ai.network/download/ for Windows, macOS, or Linux. The application includes a full node that syncs with the testnet.