Agent frameworks & infrastructure
Agent sandboxes & runtimes
1,721 agents · most starred first
Agent sandboxes & runtimes, page 13 of 18
Automatic rollback for AI agents. Python SDK for AgentRein.
Framework-agnostic transactional agent runtime with idempotent tool calls, saga compensations, and deterministic replay.
The easiest way to run AI or user generated python code safely in a docker container
CLI for AgentRun — manage AI agent infrastructure from the command line
The easiest way to run AI or user generated python code safely in a docker container. This is a fork of the original AgentRun repo with additional customization options added.
Python SDK for agents-sandbox.
Python SDK for the AgentSandbox API - sandbox execution service for Python/Bash code
AgentScale an open source control plane for generative AI agents
Agents in the cloud
A production-ready runtime framework for agent applications, providing secure sandboxed execution environments and scalable deployment solutions with multi-framework support.
A framework for hosting and scaling AI agents.
reserved
Install and configure agentsh inside AI sandbox providers
Connect your machine to AgentsHive — auto-detects CLI agents, runs as a background daemon
Agentsphere SDK that give agents cloud environments
Agentsphere Desktop
CLI tool for connecting to AgentSphere sandboxes with SSH-like functionality
Agents in the cloud
Python SDK for AgentTier — isolated, persistent Kubernetes sandboxes for humans and AI agents
A harness-engineered AI agent runtime with tool isolation, self-correction, and permission scoping
Workspace runtime and execution boundary for AI agents
SDK for deploying AI agents
SDK for AI Network P2P file sharing system
Runner for AI Agent PM - executes tasks via MCP and CLI
Universal framework for containerizing AI agents from any framework (Google ADK, AWS Strands, LangChain) into standardized Docker containers with A2A protocol support
Simple filesytem abstraction for async Python
Multi-cloud storage SDK with STS support (AWS, Aliyun OSS, Tencent COS)
An AutoGen extension that allows code execution using YepCode's secure serverless runtime
AWS Fargate code executor for AutoGen - Run code securely and scale beyond local resources
Pydantic v2 models for AWS bedrock_agentcore, shipped as a PEP 420 namespace extension of aws-resource-validator.
Pydantic v2 models for AWS bedrock_agentcore_control, shipped as a PEP 420 namespace extension of aws-resource-validator.
Backend.AI Agent
Backend.AI Storage Proxy
Headless agent daemon exposing local agent CLIs (Claude Code, Codex) over a Unix socket.
BoxLite as a local, embedded MicroVM SandboxClient for the OpenAI Agents SDK.
A multi-provider Python code execution and dependency management agent
On-premise, sandboxed AI agent platform for regulated sectors.
Run Claude Code safely inside Daytona sandboxes
CLI daemon that proxies code execution to remote Jupyter kernels with local notebook logging
Run Claude Code CLI in isolated micro-VMs with persistent workspaces
Clear Lambda code storage
Restart-on-exit supervisor for autonomous CLI agents
Interface of S3 and Google Storgae
Storage clients for different cloud providers
A unified interface for AWS S3 and Google Cloud Storage operations
CodeBuddy Cloud Agent SDK (Python) — 基于 ACP 的云端沙箱 Agent 客户端
Official Python SDK for Computer Agents. Run agents, manage cloud computers and projects, and deploy web apps, functions, databases, auth, runtimes, and secrets.
Capsule sandbox integration for CrewAI
DarrxScale Workspace Agent — remote tool execution engine for the DarrxScale Agentic System
Darwin — adaptive software systems. Includes darwin.agenticcloud, the verifiable agentic cloud. Open-source compute for AI agents with cryptographically signed attestations.
A secure sandbox execution environment for DB-GPT Agent
Docker sandbox integration for Deep Agents
Microsandbox backend for DeepAgents
Opensandbox backend for DeepAgents
Linux-native sandboxed code execution for agent workflows using bubblewrap and cgroups
A lightweight agent runtime that turns any machine into an AI-agent-ready compute target
Version 0.1.0 - Initial public release
Dragon is a composable distributed run-time for managing dynamic processes, memory, and data at scale through high-performance communication objects.
DSPy CodeInterpreter implementation using Monty, a secure Python interpreter written in Rust
Multi-language REPL engines for DSPy Recursive Language Models (RLM) with trajectory tracing and benchmark tooling
Durable execution for DSPy programs using Temporal workflows
Dataloop package runner
API for interacting with Agent
Haystack integration for E2B cloud sandboxes
The deployment layer for agentic applications. Build agents with anything. Ship them with FastAgentic.
Stateful agent runtime with tool execution and event streaming
frago - Agent OS for AI agents: runtime, resource management, environment sync, and GUI for non-technical users
Agent Dev Environment API
The OS for Recursive Agentic Intelligence
Autonomous agent runtime governed by Maelstrom Gate — eats its own dogfood
AgentOS CLI tool for managing the AgentOS Client Docker container
AgentOS CLI tool for managing the AgentOS Client Docker container
CodeInterpreterAPI is an (unofficial) open source python interface for the ChatGPT CodeInterpreter.
An Open Source version of ChatGPT Code Interpreter
A secure sandbox execution environment for GPT-DB Agent
Unified local client that grafts GPTree agent runtimes onto your machine
Dedicated runtime for Guardian helper-backed standalone skills.
Environment API for LLM agents.
A general-purpose agent harness that gives LLM agents a CLI-based computer to complete tasks the way humans do
Local-first agent OS. Spawn persistent AI agents that collaborate, write code, and use tools autonomously.
A code execution server with MCP support
init-agent-docker.
Automated Claude Code execution in isolated environments
SwarmSpawner enables JupyterHub to spawn jupyter
One-command installer for Jupyter + LLM agent sandbox
A client library to interact with the Agentic Sandbox on Kubernetes.
Library for storing and retrieving state of Kelvin applications.
Kronos Agent OS (KAOS): self-hosted runtime for durable AI agents with memory, skills, MCP tools, automations, and sub-agent coordination
Amazon Bedrock AgentCore Code Interpreter sandbox integration for Deep Agents
Deprecated: use langchain-azure-compute. Connects Azure Container Apps dynamic sessions and LangChain
Capsule sandbox integration for LangChain — run code in isolated WebAssembly sandboxes
Daytona sandbox integration for Deep Agents
LangChain tool for running Python in disposable-exec isolated containers
Docker sandbox backend for LangChain DeepAgents Harness.
Custom Docker sandbox backend for Deep Agents framework, providing isolated code execution environments
Self-hosted E2B sandbox integration for Deep Agents