converge-ai-labs/agent-foundation

a13n - Build, run, and manage AI agents at scale on your own infrastructure. Open source, with a built-in harness, durable execution, multi-provider support, and a management console. Built With Love 💜 @converge-ai-labs

What it solves

Agent Foundation (a13n) solves the challenge of running and managing AI agents reliably in production. It provides the infrastructure needed to handle long-running tasks that must persist even if a client disconnects, and ensures that interrupted processes can automatically resume from saved checkpoints if a failure occurs.

How it works

The project consists of a central service that manages agent execution, state, and configuration. You can configure agents via a web console, then integrate them into your own applications using an HTTP API or language-specific SDKs (Python, TypeScript, Go, or Rust). The system uses PostgreSQL for state management and object storage for checkpoints, allowing it to scale across multiple workers in Kubernetes or Docker environments.

Who it’s for

It is designed for developers and organizations who need to deploy AI agents on their own infrastructure. It is ideal for those who want to maintain control over their data and providers while needing robust, production-ready features like observability, usage tracking, and fault-tolerant execution.

Highlights

  • Durable execution: Automatically resumes tasks from checkpoints after worker failures.
  • Multi-language support: Includes SDKs for Python, TypeScript, Go, and Rust.
  • Built-in observability: A web console to debug tool calls, monitor runs, and track costs.
  • Flexible infrastructure: Self-hostable via Docker Compose or Kubernetes with support for custom model, memory, and sandbox providers.

Related

  • Project
  • Project
  • Project
  • Project