§ A.L.I.C.E.

Multi-Agent Orchestration Framework

The open-source framework for coordinating multiple AI agents at production scale.

UserPromptA.L.I.C.E.OrchestratorAgent-1Processing...Agent-2Processing...Agent-NProcessing...Output
§ Features

Built for production

Everything you need to coordinate agent teams at scale.

Parallel Agent Execution

Run multiple agents simultaneously with intelligent task decomposition.

Shared Memory Bus

Agents share context through a unified memory layer for coherent collaboration.

Configurable Orchestration

Define custom delegation strategies, priority rules, and convergence logic.

OpenTelemetry Observability

Built-in tracing and metrics for every agent decision and handoff.

Hot-Reload Configs

Update agent behaviors without restarting the system.

Zero-Vendor Lock-in

Runs on any LLM provider. Switch backends without changing your agent code.

§ Quick Start

Get running in minutes

alice.config.yaml
# alice.config.yaml
agents:
  - name: planner
    model: gpt-4
    role: task decomposition
  
  - name: executor  
    model: gpt-4
    role: action execution
    
  - name: reviewer
    model: gpt-4
    role: quality assurance

orchestration:
  strategy: sequential_with_feedback
  max_iterations: 5
  convergence_threshold: 0.95

Ready to orchestrate?

Start with A.L.I.C.E. today. Built on NemoClaw and OpenClaw for enterprise-grade reliability.