Skip to content

01alekseev/PetoronAI

Repository files navigation

PetoronAI

PetoronAI is a deterministic cognitive AI framework built entirely on top of PetoronHash2 (PH2) and zkPetoron-Core.

The project is designed as a local-first, privacy-first cognitive architecture focused on reasoning, planning, learning, memory, simulation, and autonomous decision making.

PetoronAI provides its own cognitive foundation that developers can use to build autonomous agents, intelligent systems, decision engines, research platforms, and future AI operating environments.


Technology Stack

  • C++20
  • PetoronHash2 (PH2)
  • zkPetoron-Core
  • CMake
  • Local-first architecture

Key Characteristics

  • Deterministic execution
  • Goal-driven cognition
  • Executive control architecture
  • Experience-based learning
  • Long-term memory
  • Knowledge management
  • World modeling
  • Strategic planning
  • Risk-aware decision making
  • Safety filtering
  • Local execution
  • Privacy by design
  • No external AI providers
  • Native Petoron ecosystem integration

Built On PetoronHash2 And zkPetoron-Core

PetoronAI is tightly integrated with:

  • PetoronHash2 (PH2)
  • zkPetoron-Core

These technologies form the cryptographic and verification foundation of the framework.


Cognitive Architecture

PetoronAI contains multiple cognitive layers.

Executive Layer

Responsible for:

  • policy selection
  • policy optimization
  • reinforcement processing
  • confidence estimation
  • evidence estimation
  • executive orchestration
  • safe action selection

Goal Layer

Provides:

  • goal management
  • goal prioritization
  • goal arbitration
  • goal adaptation
  • goal scheduling
  • goal conflict resolution

Strategy Layer

Supports:

  • strategy generation
  • strategy evaluation
  • strategy evolution
  • strategy comparison
  • strategy portfolios

Planning Layer

Supports:

  • multi-step planning
  • future action sequencing
  • objective decomposition
  • execution planning
  • long-term planning

Reasoning Layer

Provides:

  • recursive reasoning
  • counterfactual reasoning
  • hypothesis generation
  • hypothesis evaluation
  • world reasoning

Memory Layer

Supports:

  • long-term memory
  • memory retrieval
  • memory persistence
  • memory indexing
  • memory evolution

Knowledge Layer

Provides:

  • knowledge storage
  • knowledge relations
  • knowledge transitions
  • knowledge validation
  • knowledge evolution

Experience Layer

Supports:

  • experience collection
  • reinforcement learning
  • policy adaptation
  • experience statistics
  • self-improvement mechanisms

Thought Layer

Provides:

  • thought artifacts
  • thought chains
  • thought graphs
  • thought persistence
  • thought search

World Model

Supports:

  • world state tracking
  • world simulation
  • environmental reasoning
  • future projection
  • utility estimation
  • risk estimation

Current Capabilities

PetoronAI can:

  • evaluate multiple possible actions
  • compare strategies
  • estimate utility
  • estimate risk
  • learn from outcomes
  • improve policies through reinforcement
  • maintain internal memory
  • maintain a world model
  • generate and evaluate hypotheses
  • perform counterfactual analysis
  • plan future actions
  • prioritize goals
  • adapt behavior through experience

Verification And Testing

The project includes:

  • unit tests
  • integration tests
  • executive tests
  • reasoning tests
  • planning tests
  • learning tests
  • safety tests
  • reinforcement tests
  • simulation tests
  • stress tests

Current validation results:

  • 131+ automated tests passing
  • 100,000-cycle executive stress testing passing
  • Testing was conducted on a MacBook Air M1

Current Status

PetoronAI v1.0.0

  • Stable release
  • Production-ready research foundation

Intended Use Cases

PetoronAI can be used as a foundation for:

  • Autonomous agents
  • AI operating systems
  • Robotics systems
  • Research platforms
  • Simulation environments
  • Decision support systems
  • Strategic planning engines
  • Multi-agent systems
  • Privacy-preserving intelligence systems
  • Cryptographically verifiable AI systems

Architecture Overview

PetoronAI

→ Executive Layer

→ Goal Layer

→ Policy Layer

→ Strategy Layer

→ Planning Layer

→ Reasoning Layer

→ Memory Layer

→ Knowledge Layer

→ Experience Layer

→ Thought Layer

→ World Model

→ zkPetoron-Core

→ PetoronHash2


Project Vision

PetoronAI is not intended to be another language-model wrapper.

The long-term objective is the development of a deterministic cognitive architecture capable of autonomous reasoning, planning, learning, memory formation, world modeling, and verifiable intelligence built entirely on Petoron technologies.


License

PetoronAI is dual licensed:

  • PetoronAI Community License (PCL)
  • PetoronAI Commercial License (PCL-C)

Commercial use requires a valid commercial license.

PetoronHash2 and zkPetoron-Core remain independently licensed under the MIT License. The PetoronAI licenses apply solely to PetoronAI.


Commercial Licensing

PetoronAI is available under the PetoronAI Commercial License (PCL-C).

Commercial licenses are issued by the copyright holder.

Commercial license payments are currently accepted in cryptocurrency.

Supported payment methods may include:

  • ETH
  • other cryptocurrencies approved by the Licensor

Commercial license issuance, renewal, validation, and proof of existence may be recorded through privacy-preserving cryptographic workflows built on zkPetoron-Core and related Petoron technologies.

Such workflows may generate confidential attestations, proof artifacts, cryptographic commitments, and license verification records.

Where applicable, cryptographic commitments may be anchored on Ethereum through dedicated smart contracts while preserving the confidentiality of licensing details.

As a result, license ownership, license scope, payment records, and licensing metadata may remain confidential while still allowing independent cryptographic verification of license existence and authenticity.

The exact licensing process, payment methods, verification mechanisms, cryptographic attestation procedures, and smart-contract infrastructure may evolve over time and remain at the discretion of the Licensor.


Petoron | Ivan Alekseev