MAPLE v1.1.1

MAPLE v1.1.1

Multi Agent Protocol Language Engine
Creator: Mahesh Vaijainthymala Krishnamoorthy (Mahesh Vaikri)

Installation

pip install maple-oss==1.1.1

What's Changed

See CHANGELOG.md for detailed changes.

MAPLE v1.1.0 — Autonomous Agentic AI

MAPLE v1.1.0 — Autonomous Agentic AI

Multi Agent Protocol Language Engine
Creator: Mahesh Vaijainthymala Krishnamoorthy (Mahesh Vaikri)

Installation

pip install maple-oss

With LLM support:

pip install maple-oss[llm]

What's New

Autonomous Agent System

• AutonomousAgent — ReAct-loop powered agent with goal pursuit, multi-step reasoning, reflection, and backtracking
• LLM Provider Layer — Pluggable provider system supporting OpenAI, Anthropic, and compatible APIs (vLLM, Ollama, Together)
• Tool Framework — Extensible Tool and ToolRegistry with built-in MAPLE tools (send_message, query_agents, read/write state, check resources, establish links)
• Memory System — Three-tier memory: WorkingMemory (context window), EpisodicMemory (task history), SemanticMemory (learned facts)
• Multi-Agent Orchestrator — Supervisor and consensus execution patterns with capability-based team formation
• MCP Tool Discovery — Discover and register external MCP server tools as native MAPLE tools
• Observability — DecisionLogger and AgentSnapshot for full decision tracing

Infrastructure Improvements

• NATS broker auto-detection and ProductionBrokerManager wiring
• Authorization enforcement auto-initialized in MessageBroker
• Priority message queuing and routing integrated into broker delivery
• Agent auto-registration/deregistration in AgentRegistry
• Cryptographic handshake via AES-256-GCM with graceful fallback
• Consolidated circuit breaker pattern across fault tolerance and failure detection
• Built-in agent metrics (messages sent/received/failed, handler errors, processing time)

Quality

• 818 tests passing across Ubuntu, Windows, macOS
• Python 3.9 — 3.12 supported
• 80% code coverage
• Full CI/CD pipeline with automated PyPI publishing

Quick Start

from maple import AutonomousAgent, AutonomousConfig, LLMConfig, Tool

config = AutonomousConfig(
    llm=LLMConfig(provider="openai", model="gpt-4"),
    max_reasoning_steps=15
)
agent = AutonomousAgent("researcher", autonomy_config=config)
agent.register_tool(Tool(name="search", description="Search the web", handler=search_fn))
result = agent.pursue_goal("Research quantum computing breakthroughs in 2025")

Full Changelog

See CHANGELOG.md for detailed changes.

MAPLE v1.0.0 — Initial Release

MAPLE v1.0.0 — Initial Release

Multi Agent Protocol Language Engine
Creator: Mahesh Vaijainthymala Krishnamoorthy (Mahesh Vaikri)

Installation

pip install maple-oss

Core Features

• MAPLE Protocol — Multi Agent Communication Protocol with standardized message format
• Rich Type System — Comprehensive type validation with primitive, collection, and special types
• Result<T,E> Pattern — Rust-inspired error handling with explicit success/error types
• Resource Management — Resource specification, allocation, and negotiation
• Link Identification Mechanism — Secure communication channel establishment
• Distributed State Management — Consistency models for large-scale agent systems
• Communication Patterns — Request-response, publish-subscribe, streaming, and broadcast

Security

• JWT-based agent authentication
• Role-based access control
• End-to-end message encryption
• Secure link establishment

Performance

• 10,000+ agents scalability
• 5-15ms message delivery latency
• 10,000+ messages/sec throughput
• 99.99% uptime with fault tolerance

Full Changelog

See CHANGELOG.md for details.