A minimalist KVM-based virtual machine monitor that runs lightweight microVMs for secure, multi-tenant serverless and container workloads.
Agents & MCP
Sandboxes for AI agents: 34 open-source projects (2026)
Isolated environments that let AI agents run code and commands without touching your machine or data.
- Project
- 34
- This week
- +4.7K
- Updated
- 3 hours ago
How to choose
Start with what you need to isolate: a snippet of generated Python, a single process, or a full environment with a browser. If you only run AI-generated Python, pydantic/monty is a Rust sandbox that needs no containers. For hardware-isolated microVMs on your own machine, superradcompany/microsandbox runs untrusted code locally. For a full environment of code, files and browsers on your own infrastructure, look at opensandbox-group/OpenSandbox.
- montyBest for developers who only need to run AI-generated Python safely and want to avoid containers, since it is a Python sandbox written in Rust.
- microsandboxBest for anyone who wants to run untrusted code and agent workloads locally in hardware-isolated microVMs using OCI images.
- OpenSandboxBest for teams that want to give agents isolated environments for code, commands, files and browsers on their own infrastructure.
Top projects compared
| Project | Stars | This week | License | Updated |
|---|---|---|---|---|
| 37.9K | +146 | Apache-2.0 | yesterday | |
| 31.9K | +526 | MIT | 3 hours ago | |
| 23K | +76 | Apache-2.0 | 8 hours ago | |
| 21.2K | +413 | Apache-2.0 | 4 hours ago | |
| 16.5K | +159 | Apache-2.0 | 9 hours ago | |
| 14.8K | +130 | Apache-2.0 | yesterday |
All projects
Open-source computer-use toolkit giving AI agents desktops to operate, with Cua Driver, Lume VMs, cross-OS sandboxes and Cua Bench evaluation.
A reference stack from NVIDIA for running AI agents such as OpenClaw, Hermes and LangChain Deep Agents more safely inside OpenShell sandboxes.
A safe, private runtime for autonomous AI agents that enforces file, system call, and network policies per sandbox.
Open-source sandbox runtime that gives AI agents isolated environments to run code, commands, files, and browsers on your own infrastructure.
Open-source infrastructure for running AI-generated code in isolated cloud sandboxes, with JavaScript and Python SDKs.
Open-source sandbox service for AI agents that runs each workload in a hardware-isolated MicroVM, with E2B SDK compatibility and sub-60ms startup.
A virtual filesystem inside a Cloudflare Durable Object that gives AI agents a workspace with pluggable container or Worker-based execution backends.
Local microVM runtime and SDK for running untrusted code and AI agent workloads in hardware-isolated sandboxes with OCI images.
A TypeScript library that runs AI coding agents in isolated sandboxes, then merges their commits back via sandcastle.run().
A secure Python sandbox written in Rust for running code generated by AI, without containers.
Embeddable, portable microVMs in Rust for running AI agents and untrusted code locally with hardware isolation.
A lightweight CLI and library that restricts filesystem and network access for arbitrary processes at the OS level, without containers.
An all-in-one Docker sandbox giving AI agents a shared browser, shell, file system, VSCode Server, Jupyter and MCP services.
Zero-trust sandboxed runtime on Kubernetes that maps many stateful AI agents onto a small pool of workers by snapshotting idle actors.
Fork-style microVM sandboxes for AI agents: children fork copy-on-write from a warmed Firecracker parent, KVM-isolated, with live BRANCH of running VMs.
A library and hosted service that gives AI agents isolated Linux-like sandboxes with files, processes, shell, and networking, built on WebAssembly and V8.
A Kubernetes controller and CRD for managing isolated, stateful, single-pod sandboxes, aimed at AI agent runtimes and RL workloads.
Sub-millisecond KVM virtual machine sandboxes for running AI agent code, created by forking a pre-warmed Firecracker snapshot with copy-on-write memory.
Python and JS/TS SDKs for running AI-generated code in isolated cloud sandboxes and reading back its output.
A sandbox platform that runs agent and user code in isolated Firecracker microVMs, controlled through an HTTP API.
An open source Linux desktop sandbox on E2B that AI models can control through mouse, keyboard and screenshots, with a streamable view.
Python library that runs LLM-generated code in isolated sandboxes through one API across Docker, Kubernetes, Podman and local Linux or Windows backends.
An ultra-light micro-VM sandbox for running untrusted workloads with hardware-enforced isolation, built on OpenVMM with Linux as the guest.
Self-hosted Kubernetes-based infrastructure that runs untrusted code in lightweight VM sandboxes, controlled via a CLI, HTTP API and Python SDK.
A Linux-based CLI that runs AI agents in ephemeral microVMs with network allowlisting and host-side secret injection.
A runtime interface that lets AI agents run shell commands in sandboxed environments, locally or on cloud platforms like Docker, AWS, and Modal.
An open source sandbox engine for AI agents that restricts what they can execute, read, write, and reach on the network, with default-deny policies.
A sandbox wrapper that runs Claude Code (and other AI coding agents) with auto-approval inside native OS sandboxing or Docker, keeping your machine safer.
A Python library and REST API that runs LLM-generated Python code in an isolated Docker container with resource limits, timeouts, and safety checks.
FAQ
- How are these projects chosen?
- Every project on Repolode passes a quality screening before it is listed: it must be usable today, solve a concrete problem, be more than a thin wrapper and still be maintained. This page groups the listed projects by topic and ranks them by GitHub stars.
- Which project here is the most popular?
- firecracker-microvm/firecracker, with 37.9K GitHub stars. Check each project's license and last commit before choosing.