NimbleEdge AI

NimbleEdge, Inc.

About this app

Bring the power of conversational AI right into your pocket—no cloud, no servers, no data leaks.

Why NimbleEdge AI?
• Truly On‑Device: All inference runs on your phone via the Llama 1B engine, so your queries never leave your device.
• 100% Offline Mode: Chat anywhere—planes, trains, remote hikes—or simply when you want zero distractions.
• Privacy by Design: Your entire chat history stays on‑device; we never collect or store your conversations.
• Voice & Text: Talk or type— Whisper or Google ASR transcription meets the Kokoro TTS model to deliver lifelike, responsive dialogue.
• Low Latency: Instant replies without waiting for a round trip to the cloud.

How it works:
• Download & open → Downloads models for offline use
• Speak or type → Whisper or Google ASR converts speech to text in real‑time
• AI generates the reply → Llama 1B gives smart, context‑aware answers
• Kokoro TTS reads aloud → natural, human‑like voice output

Perfect for:
• Quick recipes & cocktail ideas on the fly—no internet needed
• Brainstorming & creative writing prompts in seconds
• Summarising articles, notes & long‑form content in a tap
• Drafting emails, messages & quick notes seamlessly on‑device

Download NimbleEdge AI now and unlock true on‑device intelligence—no strings attached.

Ratings and reviews

Useless. There are errors while downloading/installing the required LLM (atm you cannot use other model). Even if you achieve the LLM installation, after a while, the app might crash at starting. On top of that, the LLM model is way too simple, and makes all classic LLM errors that are uncommon in modern models. It works offline, it has voice, and its (unlimited) text generation speed is fairly fast, but it cannot access the internet. It'd interesting if the app was flawlessly, but it doesn't.

I like this concept a lot my biggest 5 star reasons are 1. No purchases and number 2. No Ads That makes me want to donate to keep it like that actually. Also love the simple interface, please make it so we can switch models and also choose wether our local LLM that we select can retain info I really would like that so I can have different LLMs to load and also so my LLM can help/understand me better by retaining the info I tell it

I really wanted to give this app a go given the developer's privacy-first intent. However many times I tried to install it, it ran to 100% then I watched the progress bar pulse ad infinitum. Three stars for putting it out there but after many attempts, never got it to work.

What’s new

* Added support for Gemini AI on Pixel devices
* Removed invite screen for a smoother onboarding experience