Home

MAPLE - Multi Agent Protocol Language Engine

Creator: Mahesh Vaijainthymala Krishnamoorthy (Mahesh Vaikri)

---

What is MAPLE?

MAPLE (Multi Agent Protocol Language Engine) is a production-grade autonomous agentic AI framework. It combines autonomous reasoning (LLM-powered ReAct loops, tool use, memory) with production infrastructure (Result<T,E>, circuit breakers, security, resource negotiation, distributed state).

v1.1.0 — The only framework with both autonomous cognition AND production-grade infrastructure.

---

Key Features

Autonomous Agentic AI

• AutonomousAgent — ReAct reasoning loop: Think, Act, Reflect with automatic backtracking
• LLM Providers — Pluggable OpenAI, Anthropic, and compatible API support (vLLM, Ollama, Together)
• Tool Framework — Extensible Tool and ToolRegistry with built-in MAPLE tools
• Memory System — Three-tier memory: Working (context), Episodic (task history), Semantic (learned facts)
• Multi-Agent Orchestrator — Supervisor and consensus patterns with capability-based team formation
• MCP Integration — Discover and register external MCP server tools as native MAPLE tools
• Observability — Full decision tracing via DecisionLogger and AgentSnapshot

Production Infrastructure

• Result<T,E> — Rust-inspired type-safe error handling
• Security — JWT authentication, RBAC authorization, AES-256-GCM encryption, secure links
• Resource Management — Specification, allocation, and multi-agent negotiation
• Circuit Breakers — Fault tolerance with automatic recovery
• Distributed State — Consistency models for large-scale agent systems
• Message Broker — Priority queuing, routing, NATS support, pub/sub, streaming

---

Quick Start

pip install maple-oss

from maple import AutonomousAgent, AutonomousConfig, LLMConfig, Tool

# Configure autonomous agent
config = AutonomousConfig(
    llm=LLMConfig(provider="openai", model="gpt-4"),
    max_reasoning_steps=15
)

# Create agent and add tools
agent = AutonomousAgent("researcher", autonomy_config=config)
agent.register_tool(Tool(
    name="analyze",
    description="Analyze data",
    handler=lambda data: Result.ok(f"Analysis: {data}")
))

# Pursue a goal — agent reasons, uses tools, reflects autonomously
result = agent.pursue_goal("Research and summarize recent AI breakthroughs")
print(result.unwrap())

---

Validation

Category	Status
Core Components	818/818 tests passed
OS Compatibility	Ubuntu, Windows, macOS
Python Versions	3.9, 3.10, 3.11, 3.12
Code Coverage	80%
CI/CD	All pipelines green
PyPI	Published as maple-oss

---

Documentation

Wiki Pages

• Installation Guide — Install from PyPI, configure extras, platform setup
• API Reference — Complete API documentation for all modules
• Autonomous Agent Guide — AutonomousAgent, tools, memory, orchestration
• Protocol Specification v1.0 — MAPLE protocol technical specification

Repository

• README — Project overview and quick start
• CHANGELOG — Version history
• Examples — Working code examples
• Documentation — Detailed guides

---

Framework Comparison

Feature	MAPLE	LangGraph	CrewAI	AutoGen
Autonomous ReAct Loop	Yes	Yes	Yes	Yes
Result<T,E> Error Handling	Yes	No	No	No
Resource Negotiation	Yes	No	No	No
Circuit Breakers	Yes	No	No	No
Secure Agent Links	Yes	No	No	No
Distributed State	Yes	No	No	No
MCP Integration	Yes	No	No	No
Multi-Agent Orchestration	Yes	Yes	Yes	Yes
Memory System	Yes	Yes	Partial	Yes
Priority Message Queuing	Yes	No	No	No

---

Community

• GitHub Issues — Report bugs and request features
• GitHub Discussions — Questions and ideas
• Releases — Download releases

---

License

MAPLE is released under the AGPL 3.0 License. See LICENSE for details.

Created by: Mahesh Vaijainthymala Krishnamoorthy (Mahesh Vaikri)