Skip to main content
Back to projects
activeJan 09, 2026

Director

A proactive meta-orchestrator that coordinates multiple AI systems and anticipates needs before you ask.

pythonaiagentslanggraphfastapipyside6desktop
GitHub

Overview

Director is the central brain that sits above other AI systems, proactively managing tasks, monitoring system health, and anticipating user needs. Unlike reactive assistants that wait for commands, Director observes patterns, predicts requirements, and takes initiative to keep everything running smoothly.

Key Features

  • 10 Specialized Agents: Orchestrator, sentry, planner, executor, memory, communication, resource manager, health monitor, learning, and reporting agents.
  • Sentry System: Continuous monitoring that detects anomalies, system degradation, and opportunities for optimization.
  • Strategy Bank: Stores successful approaches and automatically applies them to similar future situations.
  • Apocalypse DB: Disaster recovery system that maintains system state and enables rollback from failures.
  • Multi-Interface Access: Desktop GUI, CLI, and REST API for different interaction modes.
  • Proactive Scheduling: Anticipates recurring tasks and prepares resources before they're needed.

Technical Architecture

Director operates as a meta-layer above other systems, receiving telemetry and status updates while issuing high-level directives. The sentry subsystem continuously evaluates system state against expected baselines, triggering interventions when drift is detected.

Core components:

  • Orchestration Layer: LangGraph-based state machine managing agent coordination.
  • Event Bus: Async message passing between agents and external systems.
  • Decision Engine: Weighs multiple factors to determine optimal actions and timing.
  • Persistence Layer: Maintains operational history for learning and recovery.

Technology Stack

  • Core: Python 3.11+, LangGraph, LangChain
  • Backend: FastAPI for REST API and WebSocket connections
  • Desktop: PySide6 for native GUI with real-time dashboards
  • Daemon: Rust service for high-performance background operations
  • Database: PostgreSQL for operational data, Redis for caching
  • Monitoring: Custom metrics collection with Prometheus-compatible export

Current Status

Active development with core orchestration complete. The sentry system and proactive scheduling are functional. Currently improving the learning subsystem to better generalize from past successes.

Have questions about Director?

Try asking the AI assistant! Here are some ideas:

Related Projects