Atila
@atiorh
Followers
2K
Following
4K
Media
46
Statuses
486
on-device AI at @argmaxinc
San Francisco, CA
Joined July 2016
Argmax Pro SDK is now generally available!!.
Introducing Real-time Transcription with Nvidia Parakeet.- Same top accuracy as file transcription.- Best-in-market 160 ms lips-to-screen latency.- 744x more cost-efficient compared to cloud APIs.- Available in Argmax Pro SDK starting today!. Link in comments
1
3
23
I guess it is the same situation as @xeophon_ 's yapping :D Abundance FTW
the good thing about transcribing: you can yap a lot more about the relevant context. the bad thing: sometimes it transcribes wrong (like "the" -> "v") AND GPT-5 STILL TAKES THINGS TOO LITERAL.
1
0
2
I started using custom voice keyboards on iPhone to dogfood our product, and now my friends are asking me why I send them VERY long messages😅 Abundance FTW.
Developers surprise us with new use cases for background transcription on iOS!. Custom voice keyboards (like @superwhisperapp) can now dictate right in the app with local models without switching to another dedicated app. The demo video is using @NVIDIAAI Parakeet v3 with
1
0
9
RT @argmaxinc: Developers surprise us with new use cases for background transcription on iOS!. Custom voice keyboards (like @superwhisperap….
0
4
0
7W is faster and more accurate than most APIs with 700W🤔.
Speech-to-text with Argmax SDK on Mac is 40% faster now!. M1 MacBook Air + Parakeet = 226 seconds of audio transcribed in 1 second. M4 Mac mini + Parakeet = 367 seconds of audio transcribed in 1 second.
1
0
13
I had my first major collaboration with an LLM last week at the algorithmic level. It is on its way to prod, I will share the code link here once it is open-sourced next week. @OpenAI GPT5 Pro was surprisingly effective with high-level discussions as it translated Swift code to.
Continuing the journey of optimal LLM-assisted coding experience. In particular, I find that instead of narrowing in on a perfect one thing my usage is increasingly diversifying across a few workflows that I "stitch up" the pros/cons of:. Personally the bread & butter (~75%?) of.
0
0
14
I have never vented here before but @CarbonHealth just asked for a lot more money after my family’s visit was done and blocked our test results which we need urgently. @erenbali I was just trying to support a Turkish startup ðŸ˜.
2
0
4
Introducing Real-time Transcription with Nvidia Parakeet.- Same top accuracy as file transcription.- Best-in-market 160 ms lips-to-screen latency.- 744x more cost-efficient compared to cloud APIs.- Available in Argmax Pro SDK starting today!. Link in comments
0
0
1
100%. If you want to run speech-to-text and LLMs in real-time, @argmaxinc returns in 160ms so LLMs can get their turn too.
Latency is the most underrated product feature. 500ms feels instant. 1s feels broken. 2s and you’ve lost the user completely. At Wispr Flow we’ve had to rethink infra from the ground up just to hit sub-500ms LLM inference worldwide. If you like sweating the milliseconds, we’re.
1
0
9
Background transcription is required for real-world meeting notes apps because end-users transcribe hours and hours of audio without worrying about whether their phones are still recording and transcribing despite being on the lock screen.
The new Argmax Playground is out today!. Test our new background transcription feature on iOS for all-day battery life. Use Nvidia Parakeet v3 or OpenAI Whisper Large v3 Turbo. It works on the Dynamic Island as well as the Lock Screen. Link in comments
3
0
15
Excited to back this strong team! Very well-aligned with what @argmaxinc is building. @divamgupta and @ron_joshi, congrats on the launch!.
Stellon Labs (@stellon_labs) is an AI research lab building tiny frontier models that can run on any edge device. In just three weeks, they trained KittenTTS, a super-tiny speech model(<25MB), that got 8K Github Stars with 45K model downloads. Congrats
0
0
6
Macwhisper is probably the first app in the market to add the latest @NVIDIAAI Parakeet v3 speech-to-text model, ~3 days after launch. Congrats @id @jordibruin!
7
6
124
Coming to @argmaxinc Playground shortly
Major updates to Argmax Pro SDK dropped today!. - Real-time transcription in the background on iOS.- Battery-optimized mode for all-day inference and battery life.- Nvidia Parakeet v3 support in stable release. Update to 1.7.7 today! Details in comments.
2
0
15
Despite matching the Apple API in simplicity, @argmaxinc is still API-compatible with cloud transcription APIs like Deepgram:
Introducing Argmax Local Server. Run our state-of-the-art real-time transcription server directly on Mac!. 0:31 Feature complete for AI Meeting Notes apps.0:49 Migrate from cloud APIs with 1 line of code.1:05 Fastest speech models with top accuracy.1:31 Other apps do not slow
0
0
5
Apple is the best API designer in the world. @argmaxinc's new Real-time Transcription API is inspired by @Apple 's SpeechAnalyzer API.
Introducing Real-time Transcription with Nvidia Parakeet.- Same top accuracy as file transcription.- Best-in-market 160 ms lips-to-screen latency.- 744x more cost-efficient compared to cloud APIs.- Available in Argmax Pro SDK starting today!. Link in comments
2
0
22
RT @argmaxinc: Argmax @ Interspeech 2025 @ISCAInterspeech. The top-tier conference for speech technology started today! Argmax papers and e….
0
6
0