AgentScope Runtime
A production-ready runtime framework for agent apps with secure tool sandboxing, Agent-as-a-Service APIs, scalable deployment, full-stack observability, and broad framework compatibility.
What is AgentScope Runtime?
AgentScope Runtime is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to production-ready runtime framework for agent apps with secure tool sandboxing, agent-as-a-service apis, scalable deployment, full-stack observability, and broad framework compatibility.
A production-ready runtime framework for agent apps with secure tool sandboxing, Agent-as-a-Service APIs, scalable deployment, full-stack observability, and broad framework compatibility.
This server falls under the Cloud Services category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- A production-ready runtime framework for agent apps with sec
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx agentscope-runtimeConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use AgentScope Runtime
AgentScope Runtime is a production-grade Python framework that lets you deploy AI agent applications with secure tool sandboxing, Agent-as-a-Service APIs, scalable infrastructure options (local, Kubernetes, serverless), and full-stack observability. It supports multiple sandbox backends including Docker, gVisor, and boxlite, and provides specialized sandbox types for GUI, browser, filesystem, mobile, and training workloads. Developers use it to move agent prototypes into reliable production services with streaming API access and built-in logging and tracing.
Prerequisites
- Python 3.10+ and pip installed
- Docker installed and running (required for the default Docker sandbox backend)
- An MCP client such as Claude Desktop or a compatible IDE extension
- LLM provider API key for the agent framework you intend to use (e.g. Anthropic, OpenAI)
Install agentscope-runtime from PyPI
Install the core package. Add [ext] to also install optional framework integrations such as AgentScope, LangGraph, and Agno.
pip install agentscope-runtime
# or with optional framework integrations:
pip install "agentscope-runtime[ext]"Configure the sandbox backend
Set the CONTAINER_DEPLOYMENT environment variable to choose your sandbox backend. Docker is the default; gVisor and boxlite are available for stricter isolation.
export CONTAINER_DEPLOYMENT=docker
export RUNTIME_SANDBOX_IMAGE_TAG=latestCreate an Agent App
Define your agent application using the AgentApp class, decorating your query handler with the framework you want to use. The app exposes a streaming HTTP API automatically.
from agentscope_runtime.engine import AgentApp
agent_app = AgentApp(app_name="MyAgent")
@agent_app.query(framework="agentscope")
async def query_func(self, msgs, request=None, **kwargs):
# your agent logic here
pass
agent_app.run(host="0.0.0.0", port=8090)Test the Agent-as-a-Service API
Once the app is running you can send queries to it over HTTP using the /process endpoint.
curl -X POST "http://localhost:8090/process" \
-H "Content-Type: application/json" \
-d '{"input":[{"role":"user","content":[{"type":"text","text":"Hello"}]}]}'Configure your MCP client
Add the AgentScope Runtime MCP server to your MCP client configuration so AI assistants can invoke agent sandboxes as tools.
{
"mcpServers": {
"agentscope-runtime": {
"command": "npx",
"args": ["agentscope-runtime"],
"env": {
"CONTAINER_DEPLOYMENT": "docker",
"RUNTIME_SANDBOX_IMAGE_TAG": "latest"
}
}
}
}AgentScope Runtime Examples
Client configuration
MCP client JSON configuration for AgentScope Runtime with Docker sandbox backend.
{
"mcpServers": {
"agentscope-runtime": {
"command": "npx",
"args": ["agentscope-runtime"],
"env": {
"CONTAINER_DEPLOYMENT": "docker",
"RUNTIME_SANDBOX_IMAGE_TAG": "latest",
"PYTHONPATH": "/app"
}
}
}
}Prompts to try
Example prompts for interacting with AgentScope Runtime through an MCP client.
- "Run a Python sandbox and execute this data processing script"
- "Launch a browser sandbox and take a screenshot of https://example.com"
- "Start a filesystem sandbox and list the files in /tmp"
- "Execute shell commands inside a secure sandboxed environment"Troubleshooting AgentScope Runtime
Sandbox fails to start with Docker connection error
Ensure Docker is running and your user has permission to access the Docker socket. Run 'docker ps' to verify connectivity before launching the server.
CONTAINER_DEPLOYMENT environment variable not recognized
Valid values are 'docker', 'gvisor', and 'boxlite'. Check for typos and ensure the variable is set in the env block of your MCP client configuration, not just the shell.
Agent app returns 500 errors on /process requests
Check that PYTHONPATH is set to /app and that all required framework packages (e.g. agentscope) are installed in the same environment the server is running in.
Frequently Asked Questions about AgentScope Runtime
What is AgentScope Runtime?
AgentScope Runtime is a Model Context Protocol (MCP) server that production-ready runtime framework for agent apps with secure tool sandboxing, agent-as-a-service apis, scalable deployment, full-stack observability, and broad framework compatibility. It connects AI assistants to external tools and data sources through a standardized interface.
How do I install AgentScope Runtime?
Follow the installation instructions on the AgentScope Runtime GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with AgentScope Runtime?
AgentScope Runtime works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is AgentScope Runtime free to use?
Yes, AgentScope Runtime is open source and available under the Apache-2.0 license. You can use it freely in both personal and commercial projects.
AgentScope Runtime Alternatives — Similar Cloud Services Servers
Looking for alternatives to AgentScope Runtime? Here are other popular cloud services servers you can use with Claude, Cursor, and VS Code.
Open WebUI
★ 138.2kUser-friendly AI Interface (Supports Ollama, OpenAI API, ...)
Anything LLM
★ 60.4kThe all-in-one AI productivity accelerator. On device and privacy first with no annoying setup or configuration.
LocalAI
★ 46.4kLocalAI is the open-source AI engine. Run any model - LLMs, vision, voice, image, video - on any hardware. No GPU required.
Nacos
★ 33.0kan easy-to-use dynamic service discovery, configuration and service management platform for building AI cloud native applications.
Xiaozhi ESP32
★ 26.7k本项目为xiaozhi-esp32提供后端服务，帮助您快速搭建ESP32设备控制服务器。Backend service for xiaozhi-esp32, helps you quickly build an ESP32 device control server.
Gateway
★ 11.8kA blazing fast AI Gateway with integrated guardrails. Route to 1,600+ LLMs, 50+ AI Guardrails with 1 fast & friendly API.
Browse More Cloud Services MCP Servers
Explore all cloud services servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up AgentScope Runtime in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use AgentScope Runtime?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.