Stanford, CA
Joined December 2009
- Tweets234
- Following353
- Followers394
- Likes724
Pinned Tweet
Bucket list item checked ✅ Recording a cool demo video tweeted by @Google!
The incredibly powerful Gemma 4 12B is now fully rolled out globally in Eloquent. Download ai.google.dev/edge/eloquent/ and experience a powerhouse AI dictation and editing assistant that is fully local, private, subscription-free, no cap! 🧢
Our new unified architecture allows Gemma 4 12B to process multimodal inputs natively. Here's how ⬇️
Traditional models rely on separate encoders for images and audio. This adds latency and increases memory usage. So we streamlined this:
👁️ Vision: We took a novel approach to replace the encoder with a lightweight embedding module, letting the LLM backbone take over visual processing.
🎙️ Audio: We removed the encoder entirely, projecting the raw audio signal directly into the same space as text tokens.
Meet Google AI Edge Foresight, our newest Labs app (developers.google.com/edge/f…), offering an ambient note-taking meeting companion and personal knowledge assistant for Mac. Your audio, transcripts and sensitive files in your personal knowledge base never leave your device courtesy of local processing with EmbeddingGemma2 & Gemma4 suite of models.
- Live Notes Enhancement: Jot down quick shorthand notes during a meeting and Foresight enriches it in real time using the live meeting conversation.
- Ambient Audio Query Detection: Detects questions spoken aloud mid-meeting and surfaces answers live.
- Blazing-Fast Grounded Responses: Point it at your local project folders, docs, transcripts or Drive account for instant cross-modal retrieval and Q&A.
Super jazzed for the EmbeddingGemma2 launch today! It crushes the benchmarks but the real story is how much new product surface area it unlocks on the edge! As seen in the video, there's a lot to unwrap. Here is everything we built with EG2 launching today 🧵
Introducing EmbeddingGemma 2! 🚀
Our lightweight, multimodal embedding model maps text, code, images, video, and audio into a single, unified embedding space. Optimized for on-device use cases, it features:
- 740M parameter form factor with modular encoders
- Flexible dimension sizes (768dim-128dim) via Matryoshka Representation Learning (MRL)
- 8K context window (4x larger than text-only EmbeddingGemma)
- A commercially permissive Apache 2.0 license
2/ On-device MediaPipe Tasks so you can deploy EmbeddingGemma2 in your apps in a few lines of code across desktop, web and mobile! developers.google.com/edge/m…
* UniversalEmbedder: Vectorize text, images, audio, or interleaved multimodal content in one unified API.
* SemanticRetriever: An edge-optimized embedded vector search engine that pairs directly with the UniversalEmbedder for fast ANN searches in milliseconds.
3/ (On-device Jev) MediaPipe Decision API (developers.google.com/edge/m…)
EG2 is a surprisingly good decision engine! With the MP Decision API, you can define strongly typed primitives (Choice, Boolean, Score) and run zero-shot classification and routing in the order of milliseconds!
* Live Notes Enhancement: Jot down quick shorthand notes during a meeting and Foresight enriches it in real time using the live meeting conversation.
* Ambient Audio Query Detection: Detects questions spoken aloud mid-meeting and surfaces answers live.
* Blazing-Fast Grounded Responses: Point it at your local project folders, docs, transcripts or Drive account for instant cross-modal retrieval and Q&A.
ICYMI #AndroidShow ‘26 was packed with features powering by a bunch of cool @Google AI Edge tech!
Fun to see @ananyapandayy show off Screen Reactions powered by MediaPipe Segmentation (mediapipe.dev/): youtu.be/Cw-9HR6W7I8.
Another cool AI Edge / #MediaPipe powered feature is Sound Separation that helps creators turn raw footage into polished content on apps like Instagram Edits.
Also, Gemini Intelligence brings a bunch of powerful features including app automation, Gemini in Chrome, world-class AI voice dictation, and custom widgets from natural language!
Excited for all of these to roll out!
This is going to be awesome! It's amazing what Gemma4 on-device can unlock, and our #Google #AIEdge apps like Eloquent (ai.google.dev/edge/eloquent) and Gallery (github.com/google-ai-edge/ga…) are proof of that.
Super excited to announce the Gemma 4 On-Device Meetup in Bangalore! 🇮🇳💎
We're bringing the community together for an evening of hands-on building. Join us to build local agentic workflows and push the limits of on-device ML with Gemma 4 and LiteRT-LM.
Bring your laptop and let's build together. Secure your spot by RSVPing here:
luma.com/9t2074yp
Suril Shah retweeted
Gemma 4 up to 3x faster, directly in your phone! 🚀
Check out the difference Speculative Decoding makes! Multi-Token Prediction (MTP) is supercharging inference speeds for Gemma 4.
Major update to @Google AI Edge Eloquent today! 🚨 Just 2 weeks after launch, we are rolling out:
♾️ Unlimited length transcription (Most apps cap you at 5-20 mins and charge $$. We are 100% free).
📁 Audio/Video file uploads
Because it’s powered by @GoogleGemma on the AI Edge stack, it is blazing fast. ⚡
I tested it on a 32-minute, 1.8GB video of an old talk my wife gave that was buried in my phone. It transcribed and polished the whole thing in ~3 minutes! Sharing the keypoints with her from a forgotten memory was pure magic. ✨
Incredibly proud of the team for shipping an update of this scale so quickly. The reception from users has been energizing, and it reinforces our conviction: on-device AI is exactly the right approach for this space.
Try out the new update and let us know what you think
apps.apple.com/us/app/google…
@GoogleLabs
Suril Shah retweeted
‘Google AI Edge Eloquent’ is an offline, subscription-less voice dictation app 9to5mac.com/2026/04/06/googl… by @technacity
Suril Shah retweeted
Google's AI Edge Eloquent app uses AI to edit out mid-sentence mistakes to provide you with a polished transcription of your audio. pcmag.com/news/looking-for-a…
This week, we released Google AI Edge Eloquent, offering AI-enhanced dictation powered by #Gemma. The reception has been incredible, the app shot to top 20 on the iOS App Store productivity chart at its peak without any marketing.
Please try out the app and share your feedback. It’s still early days and we are going to continue making improvements to bring the benefits of #edgeAI for all. I’m proud of the team for moving fast on the user feedback and shipping near daily updates!
Spot on @outsource_! Thanks for promoting both #Google #AIEdge Eloquent and Gallery. It’s been a great reception for both the apps this week and making into top 20 of the App Store!
🚨GOOGLE QUIETLY LAUNCHED OFFLINE AI DICTATION APP
STOP PAYING FOR WISPRFLOW 🫵🏻
🤯Google just quietly dropped the future of AI dictation. AI Edge Eloquent is 100% offline
🤖Gemma-powered, and turns your messy “um… ah… wait” speech into clean, polished text instantly.
🤫No cloud. No tracking.
Beats Wispr Flow for most people.
📲iPhone (and runs on Mac). Free.
Download 👇🏻
apps.apple.com/app/google-ai…
Github 👇🏻
github.com/google-ai-edge/ga…
Great video by @omooretweets showing off #Google AI Edge Eloquent side by side with Apple dictation
