Simple wrapper for Chrome's built-in local LLM (Gemini Nano)
Chrome now includes a native on-device LLM (Gemini Nano) starting in version 138. I've been building with it since it was in origin trials, it's powerful but the official Prompt API is still a bit awkward: - Enforces sessions even for basic usage - Requires user-triggered downloads - Lacks type safety or structured error handling So I open-sourced a small TypeScript wrapper I originally built for other projects to smooth over the rough edges: github: https://github.com/kstonekuan/simple-chromium-ai npm: https://www.npmjs.com/package/simple-chromium-ai…
In plain words
Simple-chromium-ai is a TypeScript wrapper that simplifies Chrome's built-in Gemini Nano language model, available in Chrome version 138 and later. It removes friction from the official Prompt API by offering a stateless prompt method, built-in error handling with type safety, and token usage tracking. Designed for developers building on-device AI features, it provides helpers for managing model downloads while maintaining minimal overhead for rapid development and experimentation.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
Chrome now includes a native on-device LLM (Gemini Nano) starting in version 138. I've been building with it since it was in origin trials, it's powerful but the official Prompt API is still a bit awkward: - Enforces sessions even for basic usage - Requires user-triggered downloads - Lacks type safety or structured error handling So I open-sourced a small TypeScript wrapper I originally built for other projects to smooth over the rough edges: github: https://github.com/kstonekuan/simple-chromium-ai npm: https://www.npmjs.com/package/simple-chromium-ai - Stateless prompt() method inspired by Anthropic's SDK - Built-in error handling and Result-based .Safe.* variants with neverthrow - Token usage checks - Simple initialization that provides a helper to trigger downloads (must be triggered by user action) It’s intentionally minimal for hacking and prototyping. If you need fine-grained control (e.g. streaming, memory control), use the native API directly: https://developer.chrome.com/docs/ai/prompt-api Would love to hear what people build with it or any feedback!
Does a similar job
all alternatives →
- CTCode to run Gemini (Nano) locally on desktop/Chrome2024 · github.com · ▲22
Chrome Canary (nightly build for devs) now has Gemini LLM inbuilt. This is just some simple code/demo to enable and use that feature.
Open Source AI Chrome ExtensionSep 2025 · ▲87Chrome extension allow the user to communicate with the AI
- WEWebGPU enables local LLM in the browser – demo site with AI chat2025 · andreinwald.github.io · ▲145
Browser LLM demo working on JavaScript and WebGPU. WebGPU is already supported in Chrome, Safari, Firefox, iOS (v26) and Android. Demo, similar to ChatGPT https://andreinwald.github.io/browser-llm/ Code https://github.com/andreinwald/browser-llm - No need to use your OPENAI_API_KEY - its local model that runs on your device - No network requests to any API - No need to install any program - No need to download files on your device (model is cached in browser) - Site will ask before downloading large files (llm model) to browser cache - Hosted on Github…
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 19d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, July 2025
the whole month →
Trickle - Magic Canvas ▲1,064The 1st Agentic Canvas for building apps visually with AI
AI · 2025 · trickle.so
- TB
Life & fun · 2025 · trytender.app
- TY
Today marks ten years, 3653 consecutive days, of running at least one mile every day under the USRSA rules [1]. To celebrate, I built an interactive dashboard that turns a decade of GPX files into charts you can explore. Running has truly changed my life: I've made lifelong friends, explored beautiful places, and more importantly invested into my own health and fitness, which I'm starting to see the positive benefits as I get older. The stack is pretty simple: a NextJS app, with a Postgres database to keep all my running data, and all the stats are pre-computed and cached in Redis, so I…
Dev tools · 2025 · nodaysoff.run
- DA
Made this website as an exercise in vibe-coding and GCP. It was posted about a few times around the internet, on sites like Morning Brew (https://www.morningbrew.com/issues/business-buzzkill), MetaFilter (https://www.metafilter.com/209703/Draw-A-Fish), boingboing.net, etc. I think it's cute! I built a basic CNN trained against penises and swastikas, and then anything that doesn't hit the 63% confidence score gets sent to a mod queue, a vibe-coded fish-tinder…
Dev tools · 2025 · drawafish.com
