Show HN: ESP32 512kB – Tailscale, English to Python LLM and 8 containers local

Category: infrastructure

Tags: esp32, llm, embedded-ai, sandbox, tailscale, code-generation, rust

Score: 7.8/10 (Innovation: 8, Technical: 9, Documentation: 8, Utility: 6)

PySpell is a novel system that runs an LLM-powered Python-to-English translator and sandboxed code evaluator entirely on an ESP32 with 512KB RAM, reachable via Tailscale. It combines on-device AI inference, a custom IR-based code execution sandbox, and micro-container concepts for microcontrollers, showcasing impressive technical engineering.

Target audience: embedded devs, iot engineers, backend devs

Repository: https://punnerud.github.io/pyspell/ · Rust · MIT

View on Hacker News