Sandboxed Agent Execution

OpenFang.sh is an open-source agent operating system that orchestrates autonomous AI agents and capability packages across macOS, Linux, and Windows. It provides a secure, sandboxed runtime with built-in tools for tasks like research, monitoring, and automation, all managed through a native desktop

supet.sh is an agent-agnostic platform that orchestrates dozens of parallel AI coding agents in isolated git worktrees for concurrent development. It streamlines bug fixes, features, and PRs with unified workspace management, seamless IDE integration, and tools to reduce merge conflicts.

GBOX provides browser, Android, and Linux sandboxes plus RL environments for training, testing, and reproducible evaluation of agents. It includes a UI-grounding model for interface automation, iterative prototyping, debugging tools, and developer documentation.

SmythOS is an open‑source Agent Operating System that manages the AI agent lifecycle—from design to production—via visual studio, SDK, CLI, and secure sandboxed runtime. It supports multi‑platform deployment, orchestration, and enterprise‑grade security.

agentafk is an open-source unattended runtime for coding agents that enforces permission boundaries, explicit terminal states, and local trace history. It orchestrates subagents in isolated worktrees with gated operations, adversarial verification, and remote supervision for safe, auditable, long-ru

XenonStack offers a unified reasoning foundation for autonomous AI agents in operations, finance, security, and supply‑chain workflows. It supports private, edge, and multi‑cloud environments with policy‑driven governance, real‑time analytics, and seamless integration with Snowflake, Databricks, and

happycapy runs AI agents and skill modules directly in the browser, offering GUI and CLI interfaces, a skills store, browser automation, private sandboxed execution, live monitoring and action review, and developer extensibility for custom skills and workflows.

Cognitora is a cloud platform for running autonomous AI agents with managed GPUs, storage, persistent virtual machines, and vector DBs. It offers SDKs and integrations, sub-second sandboxed microVMs, auto-scaling, multi-agent coordination, live code execution and terminals.

Orca is an agent development environment for building, running, and orchestrating multi-agent workspaces with isolated worktrees, parallel agents, ghostty-style terminals, built-in editor, Git tracking, CI-aligned pipelines, observability, integrations with major LLMs, and mobile monitoring.

Reasonix.io is a terminal-native AI coding agent that leverages Deepseek’s byte-stable prefix cache to achieve 90%+ cache hits, slashing input-token costs to roughly one-fifth. It runs as a single Go binary with no Node runtime, offering persistent sessions, sandboxed plan mode, and secure local cod

paperclip is an open-source, self-hosted AI orchestration platform for creating and managing autonomous companies and agent teams—providing role-based hiring, goal-driven task delegation, budgeting, audit trails, multi-tenant deployment, extensible LLM integrations, and monitoring dashboards.

Simulation-driven platform that evaluates and monitors AI agents across modalities with realistic multi-turn scenarios, CI/CD-integrated automated tests, configurable safety/policy guardrails, and analytics for failures, hallucinations, and performance to ensure production readiness.

HoneyHive delivers AI observability and evaluation for production agents, offering OpenTelemetry tracing across 100+ LLMs, live metrics on quality, safety, latency, cost, drift alerts, offline experimentation, expert annotation, CI/CD integration, and enterprise security.

apex.ai is a comprehensive platform providing safety-certified software tools and services for autonomous systems. Its modular products enable deterministic execution, high-speed data routing, repeatable testing, and automated deployment for robotics and embedded applications.

Swarm is an experimental framework by OpenAI for orchestrating multiple AI agents in a modular, scalable manner. It enables dynamic task handoffs, function execution, and context management, making it ideal for complex, multi-agent workflows like customer support and automation.

Neo AI engineer is an autonomous agent that automates building, evaluating, and deploying ML models, LLMs, and RAG pipelines. It manages experiments, fine-tuning, and multi-step workflows, producing versioned artifacts with full evaluation and benchmarking across vendors.

VibeClaw.dev is a browser-based OpenClaw runtime for running AI coding agents in sandboxed, in-browser Node.js containers. It features a complete virtual filesystem, multi-agent orchestration, and flexible model support from local WebGPU LLMs to cloud APIs, all exportable for private, reproducible w

Buda AI is a cloud-native workspace where teams build, run, and coordinate specialized AI agents for coding, sales, marketing, and operations. It centralizes agent-driven workflows with automation, integrations, and a marketplace to reduce manual handoffs and scale routine tasks.

Maestro is an AI agent command center that orchestrates autonomous coding agents across desktop, CLI, and mobile, supporting multi-agent workflows, scheduling, Git-aware editing, remote control, real-time monitoring, and JSON/log outputs for automation.

Donovan provides a no‑code Agent Factory that builds and connects AI agents for mission‑critical government and defense workflows. It evaluates model performance, runs on classified, air‑gapped Kubernetes environments, and offers traceable reasoning with defense‑aligned guardrails.

Janus is an end-to-end simulation engine for evaluating AI agents, automating benchmark generation and creating diverse simulation environments. It enhances agent performance through continuous validation, hallucination detection, and personalized dataset evaluations.

Laminar is an open-source observability platform for AI agents that captures full trace context and browser session replay, enables rerunning and debugging executions, extracts signals and summarizes traces at scale, and provides SQL API and Evals tooling for analysis.

Verdent AI is a development environment for agentic coding that enables parallel execution of multiple agents in isolated workspaces. It streamlines workflow with context-switching, built-in clarification tools, and VS Code integration for faster, production-ready feature delivery.

Agency Swarm is an AI-powered framework that enables users to create and manage collaborative agents with specialized roles. It offers customizable agent functions, efficient communication flows, and state management, making it ideal for automating workflows and AI-driven decision-making.

Wand enables governments and enterprises to build, manage, and scale hybrid AI agent and human workforces. It offers unified oversight dashboards, SOC2‑ready security, cross‑tool interoperability, and scalable, accountable agent execution.

Tracecat is an open‑source AI automation platform for security teams, enabling workflow automation across 200+ connectors. It offers sandboxed action execution, advanced control flow, human approval gates, and case management for structured incident response.

Spine Swarm is a browser‑based platform that orchestrates multiple AI agents to execute complex tasks in parallel, delivering reports, code, dashboards or visuals without coding. It supports 300+ models and offers a no‑setup experience for non‑technical users.

OnDemand AI Agents is a decentralized OS that lets users build, deploy, and scale AI agents without a dev team. It offers a no‑code workflow builder, an agent marketplace, secure model integration, an AI playground for testing, and enterprise‑grade security.

Google Antigravity is an IDE for developers featuring tab autocompletion, natural language code commands, and configurable agents that monitor tasks. It offers cross-surface synchronization and a unified mission control view for managing multiple agents.

Trooper.AI provides private EU-hosted bare-metal GPU servers for model training, fine-tuning, and inference, with one-click AI environment templates, full root SSH and NVMe storage, tested CUDA on Ubuntu 22.04, scalable hardware and pause/upgrade controls.

Agent37 hosts and runs Claude skills on its infrastructure, letting creators upload MCP-configured skills without local setup. It provides shareable embedded links, centralized updates and versioning, analytics, onboarding, and monetization tools.

AgentX is a multi-agent AI platform for building, training, and deploying conversational agents using a no-code visual builder or developer tools, supporting multiple LLMs, RAG knowledge connectors, omnichannel deployment, integrations, analytics, voice, and on-premise options.

BlinkOps is a cloud‑native agentic security platform that automates SOC, incident response, cloud, IAM and compliance tasks through 150+ micro‑agents. Its no‑code workflow builder links to 30,000+ tools, while an AI‑as‑a‑Service team ensures configurable, auditable actions.

MuleRun is an AI agent marketplace that lets users find, customize, schedule and run agents for content creation, image generation, e-commerce, workflow automation, investment research and gaming, with one-click execution and creator publishing tools.

Fluidstack offers dedicated GPU clusters on bare‑metal Atlas OS, delivering rapid provisioning and full resource control. Continuous monitoring via Lighthouse ensures isolated, compliant infrastructure (GDPR, SOC 2, ISO 27001) with a 15‑minute support SLA for AI labs, enterprises, and government use

OpenAgents is an open-source framework for building and operating scalable, interoperable AI agent networks. It provides tools to launch, connect, and orchestrate agents with live monitoring, enabling collaborative applications and workflows.

Team9.ai is a managed OpenClaw AI-agent workspace for deploying local-first, sovereign agents as teammates. It provides zero-setup deployment, secure on-premises operation, and extensible tools for tasks like monitoring, automation, and knowledge management.

Seeko offers full‑cycle AI integration for mid‑market teams: an audit identifies high‑leverage automation, a sprint‑based program delivers production‑ready AI on the Clutch platform, and managed operations ensure ongoing optimization and compliance.

HumanLayer is an open-source IDE and orchestration layer for AI coding agents, managing parallel Claude Code sessions, multiclaude workflows, worktrees and remote workers, with context-engineering tools, session replay, workflow templates and GitHub-integrated code-review automation.

Openinterpreter is an open-source AI coding agent that runs in the terminal or as a desktop app, enabling code editing, command execution, and test automation within a sandboxed OS environment. It supports multiple model providers (OpenAI, Anthropic, local models) and features configurable permissio

Fara-7B is a 7-billion-parameter on-device AI agent that runs locally on PCs to automate desktop workflows, control applications, and execute complex task sequences for privacy-sensitive, offline system administration, data processing, and application-level automation.

Open-claw.org is a premium, subscription-based AI deployment platform that lets you launch powerful AI agents in one click

Browser Cash is an AI browser-agent platform and extension that turns browsers into secure distributed nodes, enabling sandboxed automated web tasks (research, data collection, form filling) with anonymized, isolated sessions while rewarding participants with redeemable points.

Coval lets teams test, monitor, and manage conversational AI agents by simulating thousands of realistic interactions, tracking metrics such as latency and intent accuracy, sending real‑time alerts, and supporting role‑specific workflows—all under SOC2, HIPAA, and GDPR compliance.

Runpod supplies on‑demand GPUs in 31 regions, offering single‑node pods, multi‑node clusters, and serverless workloads. It delivers low‑latency inference, efficient fine‑tuning, instant scaling, S3‑compatible storage, real‑time logs, and sub‑200 ms cold starts.

8080.ai is an AI development platform for building, orchestrating, and scaling multi-agent workflows that automate project planning, task decomposition, and sprint tracking. It provides a production-ready microservices architecture with Kubernetes deployment, a browser-based VS Code editor, and fron

AI Agent is a web app that allows users to create customized AI agents to perform specific tasks and achieve goals.

Zenfox.ai is an agentic AI platform that connects 3,000+ apps (Gmail, Slack, CRM, etc.) to execute real-system tasks and automate end-to-end workflows using natural language. It features 24/7 background agents with persistent memory, zero-code builders, and review-and-approve controls for secure, sc

Agno is an agent framework and runtime for building and running multi-agent systems, offering SDK and control plane for agents, memory/knowledge management, tool integrations, scalable cloud deployment, and security features like RBAC and request isolation.

SmolAgent is an opensource AI agent framework that simplifies creation of complex automations using pre-built llm models and agents, or custom development with open source tools.