Skip to main content

Welcome to Hephaestus

Hephaestus is a semi-structured agentic framework that enables AI agents to build workflows dynamically based on what they discover, rather than following rigid predetermined steps.

What is Hephaestus?​

Hephaestus enables you to:

  • Orchestrate Multiple AI Agents: Coordinate dozens of AI agents working in parallel on complex workflows
  • Automated Task Management: Intelligent task queuing, assignment, and lifecycle management
  • Self-Healing Monitoring: Real-time trajectory analysis with Guardian agents that detect and fix stuck agents
  • Git Worktree Isolation: Each agent works in its own isolated git workspace for conflict-free parallel execution
  • RAG-Powered Memory: Semantic memory system that enables agents to learn from each other
  • Validation Framework: Automated quality control with validator agents that verify task completion

Key Features​

🤖 Autonomous Agent Orchestration​

  • Create agents dynamically for each task
  • Intelligent task assignment based on context
  • Automatic agent lifecycle management

📊 Advanced Monitoring​

  • Guardian Agents: Monitor individual agent health and provide interventions
  • Conductor Agents: System-wide coordination and duplicate detection
  • Trajectory Analysis: Real-time behavior monitoring with LLM-powered insights

🔧 Workflow System​

  • Define multi-phase workflows in YAML or Python
  • Task dependencies and blocking relationships
  • Result validation and quality gates
  • Ticket tracking for work management

💾 Intelligent Memory​

  • Vector-based semantic search with Qdrant
  • Automatic knowledge sharing between agents
  • Memory types: errors, discoveries, decisions, learnings
  • Task deduplication to prevent duplicate work

🌳 Git Integration​

  • Worktree-based isolation for parallel development
  • Automatic branch management
  • Commit tracking and linking
  • No merge conflicts between agents

Quick Start Paths​

For New Users​

Start with our Quick Start Guide to build your first workflow in 10 minutes.

For Developers​

Dive into the Core Systems to understand the architecture, then check out the SDK Guide for programmatic control.

For Workflow Designers​

Read the Workflow Design Best Practices to learn design patterns for interconnected workflows, then explore Python Phases for advanced configurations.

Architecture Overview​

Hephaestus is built on:

  • FastAPI Server: REST API with MCP (Model Context Protocol) for Claude Code integration
  • SQLite Database: Relational data for tasks, agents, and workflows
  • Qdrant Vector Store: Semantic memory and RAG capabilities
  • tmux Sessions: Agent isolation and output capture
  • Multi-Provider LLM: Support for OpenAI and Anthropic models

Documentation Structure​

Our documentation is organized into several sections:

Getting Started​

Ready to dive in? Check out our Quick Start Guide or explore the documentation using the sidebar navigation.

Need Help? Check the Best Practices guide or explore the Core Systems for architecture details.