Skip to main content
Agent Stack is open, self-hostable infrastructure for deploying AI agents built with any framework. Hosted by the Linux Foundation and built on the Agent2Agent Protocol (A2A), it gives you everything needed to move agents from local development to shared production environments—without vendor lock-in.

​
What Agent Stack Provides

Everything you need to deploy and operate agents in production:
  • A self-hostable server that runs your agents
  • A web UI where users can test and share deployed agents
  • A CLI for deploying and managing agents
  • Built-in infrastructure services your agents can access at runtime:
    • Runtime-configurable LLMs — Switch between 15+ providers (Anthropic, OpenAI, watsonx.ai, Ollama) without changing code
    • Embeddings & vector search — Power RAG and semantic search
    • File storage — S3-compatible uploads and downloads
    • Document text extraction — Via Docling
    • External integrations — Connect to APIs, databases, Slack, Google Drive, and more via the MCP protocol with OAuth
    • Secrets management — Securely store and use API keys and credentials
  • An SDK (agentstack-sdk) that lets your agents request these services through standardized A2A extensions
  • A HELM chart for Kubernetes deployments with custom storage, databases, and auth
Build your agent using any framework—LangGraph, CrewAI, BeeAI Framework, or your own—and let the SDK handle runtime service requests automatically.

​
How It Works

Agent Stack provides the deployment layer between your agent code and production use:
  1. Build your agent using agentstack-sdk with your preferred framework
  2. Deploy with a single CLI command to your Agent Stack instance
  3. Users interact through the automatically generated web interface
Your agents request infrastructure services at runtime through the A2A protocol—no hardcoded dependencies on specific LLM providers or storage solutions. For development: Run Agent Stack locally with full services for rapid iteration For production: Deploy to Kubernetes using the provided HELM chart and integrate with your existing infrastructure (SSO, databases, cloud storage)

​
When to Use Agent Stack

Agent Stack is designed for teams who need:
  • Fast iteration across multiple agent frameworks without committing to one vendor’s hosting
  • Infrastructure control to integrate with existing corporate systems and policies
  • Internal deployment where you run the infrastructure rather than using external SaaS
  • Development velocity without rebuilding deployment infrastructure for each new agent
It complements enterprise AI platforms—use it for rapid prototyping and internal agents, while enterprise systems handle governance, security, and large-scale production workflows.

​
Get Started

Quickstart

Get up and running in one command

Connect A2A Agents

Deploy existing A2A-compatible agents to Agent Stack

Start Building Agents

Build your first agent with the Agent Stack SDK

Deploy to Production

Deploy Agent Stack to Kubernetes for your team