Production-Ready AI Agents.
Zero Code Required.

Deploy observable, secure, and scalable AI agent systems with hot reload, distributed configuration, and A2A-native federation.

Get Started GitHub
hector-server — -zsh — 80x24

A2A-Native Federation

Hector is built on the Agent-to-Agent (A2A) protocol. Unlike traditional systems with central orchestrators, Hector enables true peer-to-peer federation.

Agents communicate directly, forming federated networks where each agent maintains autonomy while collaborating seamlessly.

A2A Protocol Agent A Agent B Agent C

RAG Ready

Hector comes RAG-ready out of the box. Connect SQL databases, APIs, and document stores to build comprehensive knowledge bases.

Choose from Ollama, OpenAI, or Cohere embedders to convert your data into vectors, automatically indexed into vector databases like Qdrant. Agents retrieve relevant context using semantic search, augmenting their responses with accurate, context-aware answers.

Index SQL API Docs Vector DB

Distributed Configuration

Start simple with file-based configuration for development, then seamlessly scale to enterprise-grade distributed backends. Hector supports Consul for service mesh environments and ZooKeeper for big data infrastructure.

Enable hot reload with --config-watch for zero-downtime configuration updates. Changes are detected reactively—no polling required. Configuration updates trigger graceful reloads, keeping your agents running smoothly across all instances.

Hot Reload File Consul Zoo Keeper Config Runtime

Programmatic API

When you need to customize agent behavior, integrate with existing systems, or build dynamic agent workflows, Hector's programmatic API gives you full control.

Build agents programmatically, combine them with config-based agents, or embed agents directly into your Go applications. The configuration system is built on top of the programmatic API, meaning you can mix and match seamlessly. 

// Load agents from YAML configuration
cfg, _ := config.LoadConfig(config.LoaderOptions{
    Path: "configs/agents.yaml",
})

// Build agents from config
configBuilder, _ := hector.NewConfigAgentBuilder(cfg)
configAgents, _ := configBuilder.BuildAllAgents()

// Build a custom agent programmatically
llm, _ := hector.NewLLMProvider("openai").
    Model("gpt-4o-mini").
    APIKeyFromEnv("OPENAI_API_KEY").
    Build()

programmaticAgent, _ := hector.NewAgent("custom").
    WithName("Custom Agent").
    WithLLMProvider(llm).
    Build()

// Combine config-based and programmatic agents
runtime.NewRuntimeBuilder().
    WithAgents(configAgents).
    WithAgent(programmaticAgent).
    Start()

Everything you need for production

  • ⚡ Zero-Code Configuration

    Pure YAML agent definition. No Python/Go required—define sophisticated agents declaratively.

  • 📈 Production Observability

    Built-in Prometheus metrics and OpenTelemetry tracing. Monitor latency, token usage, costs, and errors.

  • 🛡 Security-First

    JWT auth, RBAC, and command sandboxing out of the box. Production-grade security controls.

  • 🔄 Hot Reload

    Update configurations without downtime. Reload from Consul/Etcd/ZooKeeper automatically.

  • 🔗 A2A Protocol Native

    Standards-compliant agent communication and federation for distributed, interoperable deployments.

  • 🚀 Resource Efficient

    Single 30MB binary, <100ms startup. 10-20x less resource usage than Python frameworks.