# agent-foundation (converge-ai-labs/agent-foundation) > a13n is an open-source runtime for building, running and managing AI agents on your own infrastructure. Your app calls an agent over HTTP or a Python, TypeScript, Go, or Rust SDK, and the service handles execution and recovery. - GitHub: https://github.com/converge-ai-labs/agent-foundation - GitHub.Locker page: https://github.locker/converge-ai-labs/agent-foundation - Stars: 758 - Language: Python - License: Apache-2.0 - Last push: 2026-11-07 ## Questions and answers Written by GitHub.Locker from the README and the repository; check the README below for details. **Is it a framework I code against, or a ready-made service?** It is a self-hosted service with a web Console for configuring agents. It is built on Harness, which you can also embed in your own Python code without deploying the service. **How do I get it running?** Install Docker and Docker Compose, then follow the Quickstart: create an admin account, add a model with an API key or an eligible ChatGPT subscription, create an agent and try it in Console. A Helm chart covers Kubernetes. **What do I need first?** Docker with Docker Compose for the quick start, plus a model provider API key or an eligible ChatGPT subscription. The stack runs Service, Console, PostgreSQL and Redis, and agent environments use your host Docker Engine. **Who hosts the agents, me or the vendor?** You do. It runs on your own infrastructure, and you bring your own keys for models, sandboxes, web search and memory. **Can I self-host it for free?** Yes, it is open source under the Apache 2.0 license, so you can self-host it. Your costs are your own infrastructure and provider bills. **What happens if a worker crashes mid-run?** Interrupted runs resume from saved checkpoints after worker failures, and agents keep working even when clients disconnect. **How does Harness compare to Codex?** On 62 Terminal-Bench 4.0 tasks Harness scored 58.06%, against 58.77% for published Codex results with the same model and reasoning effort, so it is roughly on par. **How do I run it for a whole team?** A Helm chart deploys it on Kubernetes, with configurable API and worker replicas and shared storage for high availability. --- ## README (checked against GitHub 2026-11-08) Source: https://github.com/converge-ai-labs/agent-foundation/blob/main/README.md

Agent Foundation (a13n)

[![CI](https://github.com/converge-ai-labs/agent-foundation/actions/workflows/ci.yml/badge.svg?branch=main)](https://github.com/converge-ai-labs/agent-foundation/actions/workflows/ci.yml) [![Documentation](https://img.shields.io/badge/docs-agent--foundation-blue)](https://a13n.dev/docs/) [![Python](https://img.shields.io/badge/python-3.13%2B-blue)](https://www.python.org/) [![License](https://img.shields.io/badge/license-Apache--2.0-green)](LICENSE) [![Slack](https://img.shields.io/badge/Slack-Join%20the%20community-4A154B)](https://join.slack.com/t/a13n/shared_invite/zt-4embv13zr-JfXjmJQop4Y9y6K3NH0kDQ) **Build, run, and manage AI agents at scale on your own infrastructure.** a13n Service is an open-source runtime for managing and running agents. Use Console, its web interface, to configure an agent, try it in a conversation, and debug its tool calls. Your application can then call that same agent through the [HTTP API](docs/a13n-service/connect-application.md) or a [Python, TypeScript, Go, or Rust SDK](docs/a13n-service/sdks.md). Service handles execution and recovery for those requests. - **Long-running tasks:** Agents keep working when clients disconnect. - **Durable execution:** Interrupted runs can resume from saved checkpoints after worker failures. - **Run visibility:** Inspect runs from your application in Console and track workspace usage and cost. Choose your [providers](#providers-and-integrations) for models, sandboxes, web search, and memory. Bring your own keys. Console and application clients connect to the Service API. Workers lease runs from PostgreSQL, execute them with Harness, and save checkpoints to object storage. After lease expiry, another attempt can resume from a saved checkpoint. Harness UI and Python apps can embed Harness directly. ## Run Service [![Console showing six managed agents, their model providers, and status](.github/assets/console-managed-agents.jpg)](.github/assets/console-managed-agents.jpg) Console with fictional demo data. See [run details from a scripted demo](.github/assets/console-run-details.jpg) and [workspace usage](.github/assets/console-usage.jpg). ### Quickstart with Docker Compose Install Docker and Docker Compose, then run: ```bash mkdir a13n-service && cd a13n-service curl -fL https://raw.githubusercontent.com/converge-ai-labs/agent-foundation/main/deploy/docker/compose/a13n-service.yaml -o a13n-service.yaml docker compose -f a13n-service.yaml up -d --wait --pull always ``` Open : 1. Create your administrator account. 2. In **Models**, select **Add model** and configure a model using an API key or an eligible [ChatGPT subscription](docs/a13n-service/models.md#chatgpt-subscription-provider). 3. Create an agent with that model and select **Try agent** to send it a message. 4. Open **Connect to app** on the same agent to get its ID, Service URL, and an API key. Call the agent from your application through the [HTTP API](docs/a13n-service/connect-application.md) or a [Python, TypeScript, Go, or Rust SDK](docs/a13n-service/sdks.md). The stack runs Service, Console, PostgreSQL, and Redis. Agent environments use your host Docker Engine. See the [Compose guide](deploy/docker/compose/README.md) for configuration. ### Production on Kubernetes Deploy to Kubernetes with our [Helm chart](deploy/kubernetes/README.md#other-clusters). Configure API and worker replicas and shared storage for [high availability](docs/a13n-service/operations.md#process-roles). ## Providers and integrations [![Bring your own stack. Self-host a13n with your choice of models, environments, web data providers, and app connections.](.github/assets/a13n-architecture.webp)](https://a13n.dev/docs/#providers-and-integrations) [Models](docs/a13n-service/models.md) · [Sandboxes](docs/a13n-service/environments.md) · [Web search](docs/a13n-service/tools.md#web-search-and-scrape) · [Memory](docs/a13n-service/memory.md) · [App connections](docs/a13n-service/tools.md#connections) ## Use Harness directly Service is built on Harness, which handles tool use, Skills, memory, and subagents. You can also use it without deploying Service: - **In your application:** [Embed Harness](docs/a13n-harness/getting-started.md) in your Python code. Your application manages execution and storage. - **On your machine:** Use [Harness UI](docs/a13n-harness-ui/index.md) to work with agents in a terminal or browser. **Benchmarks:** Harness scored **58.06%** on 62 Terminal-Bench 4.0 tasks, compared with **58.77%** in published Codex results using the same model and reasoning effort. See the [methodology](https://github.com/converge-ai-labs/agent-foundation/discussions/1015) and [FrontierHarness evaluation](https://github.com/converge-ai-labs/agent-foundation/discussions/1064). ## Documentation - [Service documentation](docs/a13n-service/index.md) - [HTTP API and application integration](docs/a13n-service/connect-application.md) - [Deployment and operations](docs/a13n-service/operations.md) - [Examples](examples/README.md) These guides track `main`. For an installed release, use its release notes and package metadata. Ask questions on [Slack](https://join.slack.com/t/a13n/shared_invite/zt-4embv13zr-JfXjmJQop4Y9y6K3NH0kDQ) or [open an issue](https://github.com/converge-ai-labs/agent-foundation/issues) for bugs and feature requests. Read [CONTRIBUTING.md](CONTRIBUTING.md) before submitting changes. You can self-host a13n under the [Apache 2.0 license](LICENSE). See [SECURITY.md](SECURITY.md) to report a vulnerability.