Enables LLMs to interact with MCP servers by writing TypeScript/JavaScript code instead of direct tool calls. Provides a code execution sandbox that accesses MCP servers through HTTP proxy endpoints.
A self-hosted MCP server that provides a single execute_code tool, enabling agents to write TypeScript to call multiple REST APIs via fetch() with transparent credential injection, reducing token usage by keeping intermediate results in the sandbox.
Code-mode MCP server (docs_search + execute_code two-tool surface) backed by a unified capability manifest across three sandbox kernels (in-process node:vm, WASM via QuickJS / Pyodide / Wasmtime, and remote microVM via E2B / Cloudflare Sandbox). At N=30 tools the bootstrap-context cost drops to 13.6% of direct tool-use. Apache-2.0.
A production-grade MCP server designed for multi-tenant, authenticated, and observable AI agent systems, enabling secure tool execution across heterogeneous data sources.
A TypeScript MCP server for launching, tracking, and managing external coding-agent runs across local and remote backends like Codex and Claude Code. It allows top-level agents to orchestrate subagents through tools for spawning tasks, polling events, and handling interactive sessions.
