esp32-ai is an experimental language-model project that runs a 28.9-million-parameter model entirely on an ESP32-S3 microcontroller. The quantized model occupies about 14.9 MB and generates text without sending data to a server. Most parameters remain in flash through a Per-Layer Embeddings design, while active computation uses SRAM and PSRAM. This memory layout allows the device to retrieve only the embedding rows required for each token. The implementation reaches roughly 9.5 tokens per second and can display generated words on a connected screen. Trained on TinyStories, the model produces short, simple stories rather than answering questions, following instructions, or providing factual knowledge. The repository includes firmware, wiring and flashing instructions, training code, quantization experiments, ablations, and measured results.

Features

  • On-device text generation
  • 28.9-million-parameter model
  • Four-bit quantized storage
  • Per-Layer Embeddings architecture
  • Flash, PSRAM, and SRAM partitioning
  • Training and firmware toolchain

Project Samples

Project Activity

See All Activity >

License

MIT License

Follow esp32-ai

esp32-ai Web Site

Other Useful Business Software
Our Free Plans just got better! | Auth0 Icon
Our Free Plans just got better! | Auth0

With up to 25k MAUs and unlimited Okta connections, our Free Plan lets you focus on what you do best—building great apps.

You asked, we delivered! Auth0 is excited to expand our Free and Paid plans to include more options so you can focus on building, deploying, and scaling applications without having to worry about your security. Auth0 now, thank yourself later.
Try free now
Rate This Project
Login To Rate This Project

User Reviews

Be the first to post a review of esp32-ai!

Additional Project Details

Programming Language

Python

Related Categories

Python Large Language Models (LLM)

Registered

2026-07-27