Agent frameworks & infrastructure
Agent sandboxes & runtimes
1,721 agents · most starred first
Agent sandboxes & runtimes, page 4 of 18
A secure file system for your agents to execute code
The Agentic-Native Operating System
Lightweight and declarative sandboxing for AI agents on Linux and macOS using Nix.
Canopy - the open-source agent workspace runtime
Composable agent runtime with enforced isolation boundaries
OpenHermit is the open-source platform for deploying fleets of AI agents as production services — durable state, sandboxed execution, managed at scale, and the channels you already use.
Self-hosted, content-addressed file hosting in one container. Upload at home, fetch by CID from anywhere. Keep your own copies and re-upload to keep links alive forever. No domain, no DNS, no cloud bill.
A framework offers an OS simulator within a Python Code Interface for AI Agents
The platform for shipping, governing AI agents — framework-agnostic, multi-runtime, GitOps-native, self-hosted.
Open-source, general-purpose sandbox platform for devs and AI agents that provides identity-based secure access to infrastructure without credentials.
Connectome agent host runtime — runs Connectome agents, MCPLs, and the context-manager stack
Programmable, verifiable execution on Xbox Series X Developer Mode, with agent-native workflows, XVM, CPU/GPU runtimes and evidence-bound execution.
Container-free RL framework for training software engineering agents
Sandbox for background agents.
A machine tool for agents: a verified runtime for systems agents build.
Aggregated Kubernetes apiserver and e2b-compatible data plane for warm-pooled agent sandboxes on microVMs. Sandboxes are served from per-node inventory with no per-sandbox etcd object and claimed from node-local sandboxd warm pools; the e2b SDK reaches them through an edge proxy.
Traverse Board · 针路簿 — 本地优先、可恢复、可审计的开源 AI Agent 工作台,面向通用代码任务与受控工作流。
Disposable Linux microVMs on your own machine, for AI agents.
通过可视化向导,根据你的开发技术栈,选择配置好Skills,MCP,Agent的AI开发工具,一键生成容器化Vibe开发环境。
A hypervisor and sandbox cloud for self-hosting AI agents and RL
AgentCgroup: Understanding and Controlling OS Resources of AI Agents
Airut is a system for running Claude Code tasks from email and Slack. It handles workspace provisioning, container isolation, network sandboxing, session persistence, and cleanup — a secure foundation for autonomous agentic development.
A pi extension with a single tool: execute, running TypeScript in a persistent Bun evaluator. Files, shell, and subagents are expressed as code rather than as more tools.
a generalized sandboxed ai-agent
Open-source AI agent runtime built in Rust. Define once, deploy isolated instances per tenant with built-in memory, encrypted vault, self-extending skills, cron, and multi-channel support.
CLI tool for running coding agents inside hardware-isolated microVMs
Lightweight WASM-based Linux VM for AI agents - run shell commands in an isolated Alpine Linux environment
Open-source sandboxes for AI agents — run untrusted, AI-generated code safely on your own machine.
⚡️ Devex — A Fast, Secure, and Scalable Repl-as-a-Service Platform built for Developers 🚀
Self-hosted platform for Claude Agent SDK agents. Container isolation, real-time execution tracking, and Web UI dashboard
The open-source AI Operating System — run a workforce of AI agents on your own machine. Any model, 1,000+ tools, a Knowledge Graph, and your own Claude Code as agents. Self-host with docker compose.
Lightweight, secure Linux sandboxes for untrusted processes. Runs in the browser and on the server.
P2P-first tunnels to private TCP/UDP services: direct QUIC with automatic relay fallback, one TCP + one UDP port for every service, built-in scoped keys and event listeners. Rust CLI plus Rust and Node SDKs, made for agent-harness remote control.
Run AI Agents in the cloud
ai agents doing agentic things on cloudflare durable objects
Starter template for deploying AI agents with Amazon Bedrock AgentCore. Complete infrastructure scaffolding with authentication, API, and web interface - all automated in one command.
Create, manage, and orchestrate stereOS AI agent sandboxes.
Sanboxes for AI agents and humans
Process manager for AI agents (and humans)
Hermes Agent (NousResearch) sandboxed by NVIDIA OpenShell — hardware-enforced network/filesystem/syscall policy, full memory + gateway stack
Local task queuing for AI agents. Prevents multiple agents from running expensive operations concurrently and thrashing your machine.
A small Recursive Language Model: let any LLM run code on its context instead of stuffing it into the prompt.
Wandesk — an open-source AI desktop. One shell, one AI kernel, one workerd; every app is a Worker site the built-in agent can write for you.
Run AI coding agents in Podman or Docker
Open source Rust framework and durable runtime for AI agents. MIT
Desktop app for managing sandboxed OpenClaw and Hermes AI agent instances via Docker. Built with Tauri v2.
High-performance, lightweight durable execution for long-running AI agents and programs.
A composable agent runtime — pair any frontend with any agent backend.
Kronos Agent OS (KAOS): self-hosted runtime for durable AI agents with memory, skills, MCP tools, automations, dashboard, and optional swarm coordination.
CLI fleet management for AI agents on local networks - deploy, upgrade, and monitor OpenClaw/ZeroClaw/IronClaw across hosts via Ansible
World-model-as-a-harness for simulating AI agent environments
Disposable Linux containers for AI coding agents, with extensible backends
The source-available AI agent operating system
A security-hardened container environment for running the OpenCode CLI in complete isolation.
面向 Android ARM64 的 Linux 桌面与本地 Agent 运行环境
High performance (LLM-) agents execution environment with production in mind.
Self-hosted execution platform for AI agents, powered by gVisor.
Agent-native streaming infrastructure. The streaming runtime built for non-human broadcasters.
AI Agent Kernel for Edge/IoT — 5-layer Rust runtime with 6 LLM providers
Full Claude agent, completely contained, one command.
Blueprint for an always-on remote agent workspace: code on your Mac, run the work on an always-on Linux box (Herdr + Syncthing + Cloudflare Tunnel + Caddy).
A lightweight microVM sandbox runtime for agents, automation, and untrusted workloads.
Run AI-generated logic safely inside your app. Stof is JSON with functions: data and logic in one document, sandboxed, in Rust, JavaScript, and Python.
Production-grade AI agent infrastructure with resource constraints, ethical governance, and cryptographic integrity (private implementation code)
Lightweight Linux sandbox for AI agents. Kernel-native isolation (namespaces, cgroups, seccomp, Landlock) with REST API, MCP bridge, and web dashboard. Single Rust binary.
dshcloud:部署并托管 DeepSeek Harness (dsh) 实例的多租户平台,支持自托管或云端部署。Deploy and host DeepSeek Harness (dsh) instances — self-hosted on your own server or cloud-hosted: a multi-tenant platform with isolated Docker containers, persistent workspaces and resource quotas.
A framework for building synthetic worlds for AI agents. Useful for RL and evaluation environments.
The enterprise-enhanced version of OpenClaw. ByClaw serves as a secure, scalable AI Agent Operating System with multi-tenancy and compliance features, providing the trusted infrastructure for large-scale organizational AI adoption.
Open-source, self-hosted control plane for OpenClaw and Hermes AI-agent fleets on Docker/Kubernetes — REST, CLI, and MCP.
Deploy your autonomous agents to production grade environments with 99% Uptime Guarantee, Infinite Scalability, and self-healing.
Rust-first, local-first multi-agent runtime with a desktop HUD, policy-gated tools, voice, and worktree-isolated coding agents.
AI coding agent sandbox platform
A curated list of code-execution sandboxing solutions for AI/LLM agents
Secure sandbox to run claude agents
Manage OpenClaw in your team (Enterprise) by providing it compute infrastructure, tool integration, Authentication and security primitives
Lightweight, Rust-native AI Agent platform. Security sandbox from day one.
OmniDaemon is a Universal Event-Driven Runtime for AI Agents, it's framework-agnostic, event-driven runtime that turns AI agents into production-grade, autonomous infrastructure services. It enables agents to listen, react, and collaborate across distributed systems—bringing true event-driven intelligence to modern enterprise environments.
AI Developer is an AI agent powered by GPT-4-Turbo that's using custom E2B Sandbox
Give your AI coding agents their own machines — not your credentials. Isolated Linux workstations with full autonomy, on Linux and macOS.
Recursive AI System
Capka AI — An open-source, self-hosted sandbox for AI agents. A community-driven alternative to Claude's Cowork
ActonOS is a single-purpose appliance operating system engineered as a customizable, self-governing AI agent kernel running 24/7.
Open-source AI agent runtime with a knowledge graph for a brain. Runs anywhere.
Self-contained ARM64 Android app for running Hermes Agent locally with a mobile interface.
An AI-native OS. Models run on your own hardware, nodes federate peer-to-peer with no broker, and the API is OpenAI-compatible. Boots, has its own Wayland compositor, and runs on 8GB. Apache 2.0.
Sandboxed Ruby for AI agents
Python Bash emulation for agents, a port of vercel-labs/just-bash
Durable agent runtime. Sessions survive crashes. Agents are just files. One binary, zero dependencies.
A computer for your agent - sandboxed code execution for AI agents
Sandboxed AI agents, reachable from anywhere
The first open-source project that turns the most powerful code agents, such as ANTHROPIC ClaudeCode and OpenAI Codex into Background Agents, accessible via REST API, a user interface, and MCP
A CLI tool for running AI agents inside microVM sandboxes
Governed infrastructure for internal tools and AI agents
Run agents in containers with credential injection and full observability.
An OpenCode plugin that sandboxes agent-executed commands using @anthropic-ai/sandbox-runtime