Side-by-side comparison of two Model Context Protocol servers — install paths, tools, quality signals, and directory engagement so you can pick the right one for Claude, Cursor, and other MCP clients.
Run untrusted Python/JavaScript code in WebAssembly sandboxes.
Control Google Colab notebooks and assign GPUs (T4/L4/A100) from any AI agent. Enhanced fork of Google's colab-mcp with all tools visible at startup, OAuth GPU control, and Windows support.
Quality signal
35/100 (Emerging)
51/100 (Fair)
Install path
pip · high
uvx · high
Engagement
0 0 0 294
0 0 0 13
Tools
Isolated execution in WebAssembly sandboxesConfigurable CPU, memory, and timeout limits per taskAutomatic retry mechanism on task failureSupports Python and TypeScript/JavaScript runtimesLifecycle tracking of running, completed, and failed tasks
Visible notebook tools at startup without waiting for browser eventsOAuth-based GPU runtime control allowing programmatic assignmentFull notebook cell lifecycle management: add, update, run, delete, moveStale server detection and cleanup to prevent orphaned connectionsWindows-compatible server port configurationOAuth token caching for persistent authorization
