Infrastructure Market: Cloud Run Sandboxes Enable Sub-Second Python in Google Apps Script

This developer article introduces a new architecture that connects Google Apps Script (GAS) with Google Cloud Run Sandboxes, which use gVisor micro-virtualization, to enable deterministic sub-second execution of Python and Bash scripts directly from Google Workspace. The solution overcomes GAS's standard limits (6-minute timeout, V8 JS only) by offloading compute to a Cloud Run Gen2 service that runs code in an isolated gVisor sandbox. The author provides an open-source implementation (GitHub repository) and documents an 8-axis test suite that verifies security properties like SSRF protection, environment variable isolation, read-only filesystem, and network egress blocking. Practical use cases include rendering Seaborn heatmaps from spreadsheet data directly into Google Sheets, with latencies of 200-450 ms. The article also compares this approach to Gemini Managed Agents, highlighting trade-offs in latency, cost, and persistence. The architecture leverages Google Cloud's free tier to minimize costs, with scale-to-zero when idle.