#1010th of 24 in Desktop AI apps

gallery

Mobile app for running open LLMs offline on Android and iOS

Stars
25.2k
License
Apache-2.0
Last commit
Oct 2026
Last release
Oct 2026
Language
Kotlin
Stars, last 28 weeks+4,007
18 Apr 2026: 21,166 stars25 Apr 2026: 21,723 stars2 May 2026: 22,169 stars9 May 2026: 22,549 stars16 May 2026: 22,730 stars23 May 2026: 22,941 stars30 May 2026: 23,083 stars6 Jun 2026: 23,337 stars13 Jun 2026: 23,472 stars20 Jun 2026: 23,604 stars27 Jun 2026: 23,691 stars4 Jul 2026: 23,784 stars11 Jul 2026: 23,886 stars18 Jul 2026: 24,018 stars25 Jul 2026: 24,090 stars1 Aug 2026: 24,162 stars8 Aug 2026: 24,240 stars15 Aug 2026: 24,323 stars22 Aug 2026: 24,399 stars29 Aug 2026: 24,465 stars5 Sept 2026: 24,531 stars12 Sept 2026: 24,593 stars19 Sept 2026: 24,660 stars26 Sept 2026: 24,717 stars3 Oct 2026: 24,778 stars10 Oct 2026: 24,874 stars17 Oct 2026: 24,971 stars24 Oct 2026: 25,092 stars31 Oct 2026: 25,173 stars

21,166 on 18 Apr 202625,173 on 31 Oct 2026

From GitHub's daily star counts.

Checked on GitHub.

Overview

Google AI Edge Gallery is a Kotlin app for Android and iOS, with a macOS download, that runs open models on-device through LiteRT. It bundles chat with speech input and output, image questions, audio transcription, prompt testing, agent skills and on-device semantic media search. Models come from Hugging Face or are loaded as custom files, and can be benchmarked on the device.

Who it is for: Mobile users and developers testing on-device LLMs on phones

Strengths

  • Inference runs on-device via LiteRT, so prompts and media stay offline
  • Chat, image Q&A, audio transcription, prompt lab and media search in one app
  • Built-in benchmarks show how each model performs on your own hardware
  • Loads custom models, pulls from Hugging Face; Agent Skills can load from a URL

Weaknesses

  • README calls it an experimental Beta release
  • Requires Android 12+ or iOS 17+; the macOS build is versioned 0.1.0
  • No server or API exposed; usable only through the app UI
  • Model choice is limited to formats LiteRT supports

What it needs

  • Needs LiteRT, Hugging Face (model download)
  • Models: LiteRT models, EmbeddingGemma 2, FunctionGemma 270m finetunes, custom models

Quick answers

Is gallery open source?

Yes. Its code is on GitHub under the Apache-2.0 license.

Also in Desktop AI apps

See all 24
Also in Desktop AI apps
RankProjectScore
1cherry-studioDesktop client for chatting with many LLM providers, with agents and knowledge bases53.2k starsAGPL-3.092 out of 100
2dsh-desktopNative Windows and macOS desktop client for DeepSeek Harness, built on plugins34.4k starsMIT88 out of 100
3screenpipeRecords screen and audio locally, makes it searchable for you and your agents22.2k starscustom license88 out of 100
#4prompt-optimizerPrompt rewriting and testing tool for web, desktop, Chrome and MCP Live demo ↗ (opens in a new tab)40.2k starscustom license85 out of 100
#5deep-live-camReal-time webcam face swap and video deepfake from a single source image97.5k starsAGPL-3.084 out of 100