Fountain
An open-source conversational API that enables autonomous AI agents to execute coding work in isolated sandboxes with access to repositories, tools, and credentials. Self-hosted platform for agent orchestration with streaming output and multi-turn convers
🔗 Visit Fountain
Description
Giving an AI agent the ability to actually run code — clone a repo, install packages, use your credentials, execute commands — is powerful but risky if you're handing that access to a third-party cloud service you don't control. Fountain's pitch is "your server, your database, your keys, your agents": a way to give an app a computer it can talk to, without giving up ownership of the infrastructure or the data that flows through it.
Fountain is an open-source (AGPL-3.0-or-later), self-hosted conversational API that lets AI agents execute work in isolated sandbox environments with access to repositories, tools, and credentials, streamed back over REST, SSE, or webhooks. It supports reusable machine configurations, an integrated credential vault, OpenAI-compatible chat protocol, the Editor Context Protocol (ACP), MCP integrations, and prebuilt connectors for 30+ external services (Anthropic, OpenAI, Google Gemini, GitHub, Slack, Stripe, and others), serving use cases from visible developer workbenches to fully automated, hidden webhook integrations.
💬 Our review
The short version: Fountain's differentiator against cloud sandbox APIs is entirely about control — self-hosting means your code execution, credentials, and conversation data never leave infrastructure you own.
E2B is the best-known cloud sandbox API for AI agents and is easier to get started with since there's nothing to self-host, but that convenience means trusting a third party with the sandbox environment and, depending on setup, the data flowing through it. Fountain trades that convenience for sovereignty: AGPL-3.0 licensing, a self-hosted credential vault, and full infrastructure ownership, which matters a lot for teams with data-residency or security requirements but means you're also responsible for operating and securing the deployment yourself. It's early-stage (a single OSS launch post, not yet a mature ecosystem), so expect rougher edges and a smaller community than an established cloud provider — worth adopting now if data sovereignty is a hard requirement, otherwise a hosted sandbox API will get you running faster.
💰 Pricing
📊 Global score
🤖 AI-enriched data
Gratuit, licence AGPL-3.0-or-later, auto-hébergé, aucune offre cloud managée mentionnée.
Pros
Auto-hébergé, souveraineté totale sur les données et identifiants
Coffre-fort d'identifiants intégré
Compatible protocole de chat OpenAI, MCP et ACP
Plus de 30 connecteurs prêts à l'emploi (GitHub, Slack, Stripe...)
Gratuit et open source (AGPL-3.0)
Cons
Projet en phase précoce, écosystème et communauté encore restreints
Responsabilité d'exploitation et de sécurisation à la charge de l'utilisateur
Pas d'offre cloud managée pour éviter l'auto-hébergement