i-am-bee/agentstack

Deploy, and share agents with open infrastructure, free from vendor lock-in.

What it solves

Agent Stack provides a standardized way to turn AI agents into scalable, running services. It removes the need for developers to build custom deployment infrastructure from scratch when moving an agent from a local prototype to a production-ready HTTP service.

How it works

It acts as an open infrastructure layer that wraps existing agents—regardless of the framework used to build them (such as LangGraph or CrewAI)—and exposes them as services via the Agent2Agent (A2A) Protocol. The stack includes a self-hostable runtime, LLM routing for 15+ providers, vector storage for RAG, and a CLI for management. It also integrates with the Model Context Protocol (MCP) for external tool access and provides a Helm chart for Kubernetes deployment.

Who it’s for

Developers building agent-powered features who want a fast path to deployment without being locked into a proprietary vendor's ecosystem.

Highlights

  • Framework Interoperability: Supports agents built with LangGraph, CrewAI, or custom frameworks.
  • Production-Ready Tooling: Includes S3-compatible file storage, secrets management, and OAuth support for secure integrations.
  • Deployment Flexibility: Offers a fast local development loop and a Kubernetes-ready architecture via Helm charts.
  • Extensive Reference Agents: Provides pre-built examples for RAG, vulnerability scanning, flight search, and multi-turn conversational assistants.

Related

  • Project
    agentscope-ai/agentscope

    A production-ready agent framework that provides an SDK for composing AI agents and a FastAPI-based service for deploying them as multi-tenant applications.

  • Project
    a2aproject/A2A

    An open protocol that enables communication and interoperability between AI agents built on different frameworks and servers, allowing them to collaborate without exposing internal logic.

  • Project
    Haohao-end/openagent

    A full-stack AI agent platform for building, orchestrating, and publishing AI applications with visual workflows, RAG integration, and OpenAPI delivery.

  • Project
    deepset-ai/haystack

    An open-source AI orchestration framework for building production-ready RAG systems, autonomous agents, and LLM applications in Python.

  • Project
    diggerhq/opencomputer

    A framework for building and deploying serverless AI agents as code, featuring real-time cloud sync and an integrated React frontend for interaction.