Feature
MCP Integration
Native Model Context Protocol support for AI agents and LLMs. Give Claude, GPT, and other AI assistants the ability to execute code safely.
mcp_config.json
1# MCP Server Configuration (cursor settings)
2{
3 "mcpServers": {
4 "hopx-sandbox": {
5 "command": "npx",
6 "args": ["-y", "@anthropic/hopx-mcp-server"],
7 "env": {
8 "HOPX_API_KEY": "your-api-key"
9 }
10 }
11 }
12}
13
14# Available MCP Tools:
15# - create_sandbox: Create isolated execution environment
16# - run_code: Execute code in sandbox
17# - run_command: Run shell commands
18# - read_file: Read files from sandbox
19# - write_file: Write files to sandbox
20# - list_files: List directory contents
21# - kill_sandbox: Terminate sandbox
22
23# Claude/AI can now safely:
24# - Write and test code
25# - Install packages
26# - Process files
27# - Run scripts
28# All in isolated sandboxes!AI Agent Ready
Native integration for Claude, GPT, and other LLMs
Isolated Execution
Run agent tools safely in isolated sandboxes
Fast Startup
~100ms sandbox creation for responsive agents
Tool Ecosystem
Pre-built tools and easy custom tool creation