Weekend PoC: Meet J-E-F-E 🤖 I put an Android phone on a Thymio II robot: on-device face detection (ML Kit), whisper.cpp ears, local reflexes in 0.5 s, an LLM brain for the rest. Plus a SpaceX-style control tower. 🤖
#Thymio #robotics #MLKit #whispercpp
#Thymio #robotics #MLKit #whispercpp
October 31, 2026 at 3:40 AM
Weekend PoC: Meet J-E-F-E 🤖 I put an Android phone on a Thymio II robot: on-device face detection (ML Kit), whisper.cpp ears, local reflexes in 0.5 s, an LLM brain for the rest. Plus a SpaceX-style control tower. 🤖
#Thymio #robotics #MLKit #whispercpp
#Thymio #robotics #MLKit #whispercpp
デフォルトのMLKitによる低精度な翻訳に加えて、Gemini Nanoによる翻訳およびGoogle翻訳へのリダイレクト機能を追加
Material Youへの対応
AndroidでもKawaii modeを使用可能に
ブロックの回避
アカウント作成日時などのアカウント詳細を確認できる機能
他人がブロックしたユーザーの確認
自分の投稿を削除して再編集する機能
確かこれぐらいだったはず
Material Youへの対応
AndroidでもKawaii modeを使用可能に
ブロックの回避
アカウント作成日時などのアカウント詳細を確認できる機能
他人がブロックしたユーザーの確認
自分の投稿を削除して再編集する機能
確かこれぐらいだったはず
個人的に弄り回したBlueskyのapk、discord鯖の雑談チャンネルに投げたので勝手に使っていいよ
October 31, 2026 at 11:25 AM
デフォルトのMLKitによる低精度な翻訳に加えて、Gemini Nanoによる翻訳およびGoogle翻訳へのリダイレクト機能を追加
Material Youへの対応
AndroidでもKawaii modeを使用可能に
ブロックの回避
アカウント作成日時などのアカウント詳細を確認できる機能
他人がブロックしたユーザーの確認
自分の投稿を削除して再編集する機能
確かこれぐらいだったはず
Material Youへの対応
AndroidでもKawaii modeを使用可能に
ブロックの回避
アカウント作成日時などのアカウント詳細を確認できる機能
他人がブロックしたユーザーの確認
自分の投稿を削除して再編集する機能
確かこれぐらいだったはず
Flutter OCR on Android: bundled ML Kit adds about 4 MB per script per architecture. Unbundled adds about 260 KB and downloads its model first.
boffincoders.com/blog/flutter...
#flutter #mlkit
boffincoders.com/blog/flutter...
#flutter #mlkit
October 30, 2026 at 6:02 PM
Flutter OCR on Android: bundled ML Kit adds about 4 MB per script per architecture. Unbundled adds about 260 KB and downloads its model first.
boffincoders.com/blog/flutter...
#flutter #mlkit
boffincoders.com/blog/flutter...
#flutter #mlkit
🚀 Android Skills v1.0.18 is here!
🔥 4 new skills + 15 updated
🤖 On-device AI with Gemini Nano
✍️ Jetpack Ink drawing & custom brushes
📱 Improved AppWidget management
Explore all 36 skills 👇
github.com/android/skil...
#AndroidDev #AI
🔥 4 new skills + 15 updated
🤖 On-device AI with Gemini Nano
✍️ Jetpack Ink drawing & custom brushes
📱 Improved AppWidget management
Explore all 36 skills 👇
github.com/android/skil...
#AndroidDev #AI
Release Release v1.0.18 · android/skills
Release v1.0.18 of Android Skills.
New skills (4)
mlkit-agent-api (device-ai/mlkit-agent-api) — Builds autonomous on-device agentic workflows in Android apps using the ML Kit GenAI Agent API power...
github.com
October 27, 2026 at 4:44 AM
🚀 Android Skills v1.0.18 is here!
🔥 4 new skills + 15 updated
🤖 On-device AI with Gemini Nano
✍️ Jetpack Ink drawing & custom brushes
📱 Improved AppWidget management
Explore all 36 skills 👇
github.com/android/skil...
#AndroidDev #AI
🔥 4 new skills + 15 updated
🤖 On-device AI with Gemini Nano
✍️ Jetpack Ink drawing & custom brushes
📱 Improved AppWidget management
Explore all 36 skills 👇
github.com/android/skil...
#AndroidDev #AI
Afaik it uses Google's MLKit on Android, which is free and on-device. People wanted in-line translations on the app instead of being redirected to an app/website and using the Google Translate API should be pretty expensive so I guess that was the "best" solution.
August 29, 2026 at 8:43 PM
Afaik it uses Google's MLKit on Android, which is free and on-device. People wanted in-line translations on the app instead of being redirected to an app/website and using the Google Translate API should be pretty expensive so I guess that was the "best" solution.
Extremely minimal Compose Multiplatform sample that demonstrates use of on-device AI on iOS and Android. github.com/joreilly/OnD... #CMP #KMP #AI
Just basic initial implementation so far that uses the following:
👉 Foundation Models Framework (iOS)
👉 MLKit Prompt API (Android)
Just basic initial implementation so far that uses the following:
👉 Foundation Models Framework (iOS)
👉 MLKit Prompt API (Android)
GitHub - joreilly/OnDeviceAI: Extremely minimal Compose Multiplatform sample that demonstrates use of on-device AI on iOS and Android.
Extremely minimal Compose Multiplatform sample that demonstrates use of on-device AI on iOS and Android. - joreilly/OnDeviceAI
github.com
November 23, 2025 at 4:21 PM
Extremely minimal Compose Multiplatform sample that demonstrates use of on-device AI on iOS and Android. github.com/joreilly/OnD... #CMP #KMP #AI
Just basic initial implementation so far that uses the following:
👉 Foundation Models Framework (iOS)
👉 MLKit Prompt API (Android)
Just basic initial implementation so far that uses the following:
👉 Foundation Models Framework (iOS)
👉 MLKit Prompt API (Android)
Flutter has cross-platform support for MLKit but I'm guessing bluesky isn't using Flutter 😅
September 3, 2024 at 5:07 PM
Flutter has cross-platform support for MLKit but I'm guessing bluesky isn't using Flutter 😅
otherwise you gotta install gapps on any aosp derived android to take advantage of mlkit or litert shit
March 10, 2026 at 6:14 AM
otherwise you gotta install gapps on any aosp derived android to take advantage of mlkit or litert shit
Possibly dumb question: can you use MLKit from AOSP derivatives?
November 30, 2025 at 3:43 PM
Possibly dumb question: can you use MLKit from AOSP derivatives?
Want to add text recognition to your Android app? 🧠
This free article walks through using ML Kit to detect text from images — all on-device.
📱 kodeco.com/47836688
#AndroidDev #MLKit #MachineLearning
This free article walks through using ML Kit to detect text from images — all on-device.
📱 kodeco.com/47836688
#AndroidDev #MLKit #MachineLearning
July 18, 2025 at 1:02 PM
Want to add text recognition to your Android app? 🧠
This free article walks through using ML Kit to detect text from images — all on-device.
📱 kodeco.com/47836688
#AndroidDev #MLKit #MachineLearning
This free article walks through using ML Kit to detect text from images — all on-device.
📱 kodeco.com/47836688
#AndroidDev #MLKit #MachineLearning
Google Opens Gemini Nano On-Device AI to Android App Developers
#AI #Google #GoogleGemini #GeminiNano #AndroidDev #OnDeviceAI #MLKit #MobileAI #GoogleIO #Privacy #Android #Devs
winbuzzer.com/2025/05/19/g...
#AI #Google #GoogleGemini #GeminiNano #AndroidDev #OnDeviceAI #MLKit #MobileAI #GoogleIO #Privacy #Android #Devs
winbuzzer.com/2025/05/19/g...
Google Opens Gemini Nano On-Device AI to Android App Developers - WinBuzzer
Google is expanding Gemini Nano access to third-party Android developers via new ML Kit GenAI APIs, enabling on-device AI features like summarization and image description for enhanced privacy and offline...
winbuzzer.com
May 19, 2025 at 10:24 AM
Google Opens Gemini Nano On-Device AI to Android App Developers
#AI #Google #GoogleGemini #GeminiNano #AndroidDev #OnDeviceAI #MLKit #MobileAI #GoogleIO #Privacy #Android #Devs
winbuzzer.com/2025/05/19/g...
#AI #Google #GoogleGemini #GeminiNano #AndroidDev #OnDeviceAI #MLKit #MobileAI #GoogleIO #Privacy #Android #Devs
winbuzzer.com/2025/05/19/g...
Uses the Apple Translation framework on iOS and Google MLKit on Android. Needs language models downloaded on the device. MLKit does this automatically in the background, iOS requires you to download the models in the Translate app.
June 19, 2026 at 7:20 AM
Uses the Apple Translation framework on iOS and Google MLKit on Android. Needs language models downloaded on the device. MLKit does this automatically in the background, iOS requires you to download the models in the Translate app.
Merged PR into main at bluesky-social/expo-translate-text: Bump mlkit language-id to 17.0.6 for 16kb page alignment (#6)
March 4, 2026 at 12:07 PM
Merged PR into main at bluesky-social/expo-translate-text: Bump mlkit language-id to 17.0.6 for 16kb page alignment (#6)
Wouldn't it be nice to have an example of correctly using #CameraX for #firebase #MLKit to showcase barcode, QRcode, face-detection et al scanning?
This would eliminate 1000+ lines of boilerplate code required to manage the camera https://t.co/aTRcVtYoz0
#AskAndroid #AndroidDev
This would eliminate 1000+ lines of boilerplate code required to manage the camera https://t.co/aTRcVtYoz0
#AskAndroid #AndroidDev
quickstart-android/mlkit at master · firebase/quickstart-android
Firebase Quickstart Samples for Android. Contribute to firebase/quickstart-android development by creating an account on GitHub.
github.com
December 1, 2024 at 10:08 PM
Wouldn't it be nice to have an example of correctly using #CameraX for #firebase #MLKit to showcase barcode, QRcode, face-detection et al scanning?
This would eliminate 1000+ lines of boilerplate code required to manage the camera https://t.co/aTRcVtYoz0
#AskAndroid #AndroidDev
This would eliminate 1000+ lines of boilerplate code required to manage the camera https://t.co/aTRcVtYoz0
#AskAndroid #AndroidDev
I want to try to deep dive into MLKit GenAI, I'm really curious how limited it would be and how would it be possible to still use the "good enough" on-device capabilities and just augment it?
January 26, 2026 at 10:52 PM
I want to try to deep dive into MLKit GenAI, I'm really curious how limited it would be and how would it be possible to still use the "good enough" on-device capabilities and just augment it?
It's kimdnof hard to explore some of the MLKit stuff, at least I'm the wild because of the small number of devices capable of running the GenAI APIs.
January 26, 2026 at 8:59 PM
It's kimdnof hard to explore some of the MLKit stuff, at least I'm the wild because of the small number of devices capable of running the GenAI APIs.
I was looking into how to program the NPU in the Apple silicon chips and there’s simply no way to do it directly, you HAVE to go through their MLKit abstractions
How am I supposed to do anything useful with it if it’s so locked down?
How am I supposed to do anything useful with it if it’s so locked down?
July 5, 2025 at 9:36 AM
I was looking into how to program the NPU in the Apple silicon chips and there’s simply no way to do it directly, you HAVE to go through their MLKit abstractions
How am I supposed to do anything useful with it if it’s so locked down?
How am I supposed to do anything useful with it if it’s so locked down?
fyi Google MLkit
November 14, 2025 at 9:29 AM
fyi Google MLkit
Easily found the biggest offender… MLKit Entity Extraction.
- 1.8MB of native for a single architecture
- 1.5MB of dex
And this doesn't include any model downloaded on demand (probably a few more MB) 😱
- 1.8MB of native for a single architecture
- 1.5MB of dex
And this doesn't include any model downloaded on demand (probably a few more MB) 😱
January 17, 2025 at 1:34 PM
Easily found the biggest offender… MLKit Entity Extraction.
- 1.8MB of native for a single architecture
- 1.5MB of dex
And this doesn't include any model downloaded on demand (probably a few more MB) 😱
- 1.8MB of native for a single architecture
- 1.5MB of dex
And this doesn't include any model downloaded on demand (probably a few more MB) 😱
github.com/bluesky-soci... we have a PR for it, unfortunately it hugely increases the bundle size so we're unsure if it's the right move
Add android on-device translation by fobos531 · Pull Request #4312 · bluesky-social/social-app
Builds on top of #4098
This PR aims to add on-device translation to Android via MLKit's Translation API to achieve feature parity with iOS. From my testing, Google's 1st party service (tran...
github.com
September 3, 2024 at 3:23 PM
github.com/bluesky-soci... we have a PR for it, unfortunately it hugely increases the bundle size so we're unsure if it's the right move
we'd also be able to do this on Android as well - we would need to add mlkit i think which is a fairly chunky dependency though it has some other uses like on-device translation which would be handy (we're shipping native iOS translation support this week as well instead of opening in google)
June 5, 2024 at 5:04 AM
we'd also be able to do this on Android as well - we would need to add mlkit i think which is a fairly chunky dependency though it has some other uses like on-device translation which would be handy (we're shipping native iOS translation support this week as well instead of opening in google)
yea. my main issue with using mlkit is you also have to include a google "analytics" library for it to function, which just isn't cool.
it also increases app bundle size significantly, but that isn't the big blocker to me as the phoning home.
it also increases app bundle size significantly, but that isn't the big blocker to me as the phoning home.
February 18, 2024 at 2:30 AM
yea. my main issue with using mlkit is you also have to include a google "analytics" library for it to function, which just isn't cool.
it also increases app bundle size significantly, but that isn't the big blocker to me as the phoning home.
it also increases app bundle size significantly, but that isn't the big blocker to me as the phoning home.