IzwiAI retweeted
Introducing Muse Wonder, our new family of frontier models with state of the art performance for personal and professional agentic work.
IzwiAI retweeted
When several people talk at once, a transcript can get messy fast.
Our new Nemotron 3 Diarization model tracks who spoke when, even when voices overlap. It handles up to eight speakers, has 100M parameters, and is now available on @huggingface 🤗
IzwiAI retweeted
2026 feels very different from previous years.
Single-GPU optimization is becoming increasingly commoditized: profile the timeline, optimize hotspots, adopt TRT-LLM/FlashInfer kernels, have agents write kernels, or borrow ideas across engines. Spec(EAGLE-3/DSpark/DFlash2) and quant(NVFP4 with ModelOpt) are all becoming increasingly standardized.
The real differentiation is moving to system-level optimization: PD disaggregation and KV stores — maximizing TPM/GPU under TPS SLAs while pushing cache hit rates higher.
Engine optimization is becoming increasingly deterministic and programmatic.
But there are still a few months left in 2026. We’ll still see plenty of interesting work ahead — new models, new hardware, and new ideas on existing ones. Looking forward to it!
IzwiAI retweeted
That’s what we are doing @IzwiAI, building voice models that understand African languages.
Local first voice AI for speech, chat, and audio workflows.
github.com/izwi-ai/izwi
IzwiAI retweeted
We promised open weights for Qwen3.8. Now, time to meet them! 🎉
⚡ Qwen3.8-27B:
- A native multimodal dense model. With just 27B parameters, it outperforms Qwen3.7-Plus overall and shines in real-world coding & office workflows.
- 262K native context, easily extendable to 1M tokens via YaRN.
- Built for builders. Highly efficient, high-quality, and licensed under Apache 2.0.
🚀 The open weights for Qwen3.8-2.4T-A95B (Max-level) have also been released recently.
Whether you're shipping lightweight applications with Qwen3.8-27B locally or building agents with Qwen3.8-2.4T-A95B, they're yours now!
Download, deploy, and build something we haven't imagined yet. 👀👇
- Hugging Face：
huggingface.co/collections/Q…
- ModelScope：
modelscope.cn/collections/Qw…
IzwiAI retweeted
GPT 5.6 Sol is truly amazing. While working on @IzwiAI, a voice AI runtime, Sol completely understands the task and just does it right the first time. The code has gotten a lot more stable in the last week 🚀
IzwiAI retweeted
IBM’s Granite Speech 4.1 2B Plus is built for rich transcription: speaker labels, word timestamps, keyword biasing.
Best way to test it locally: @IzwiAI.
`izwi pull Granite-Speech-4.1-2B-Plus`
Local-first voice AI stack. OpenAI-style APIs.
github.com/izwi-ai/izwi
IzwiAI retweeted
NVIDIA just dropped Nemotron 3.5 ASR Streaming 0.6B, small, fast, multilingual ASR for real-time voice agents.
Best way to kick the tires locally: @IzwiAI.
Run it local-first, test APIs, swap models, ship faster.
github.com/izwi-ai/izwi
IzwiAI retweeted
Izwi is the voice of AI
I've been heads down working on @IzwiAI, making it more stable and faster, and adding more models.
v0.1.0-beta-17 is out!!!
github.com/izwi-ai/izwi/rele…
#LocalAI #ASR #OpenSource
IzwiAI retweeted
Voice AI runtime. Local first transcription, speaker diarization, TTS, and voice cloning with an OpenAI compatible API.
@IzwiAI is bringing Voice AI to your computers.
Voice AI runtime. Local first transcription, speaker diarization, TTS, and voice cloning with an OpenAI compatible API.
Bringing Voice AI to your computers.
IzwiAI retweeted
Building infrastructure without VC means consulting to survive. Every hour consulting is an hour not building.
Funded competitors sprint ahead while you're stuck in the consult-build loop.
Wouldn't have it any other way.
Heads down, back to building @IzwiAI.
IzwiAI retweeted
Every time I pitch @IzwiAI, people think I've lost my mind. "How do you compete with billion dollar labs as a solo founder?"
Simple: They are building for their cloud. I'm building for your local device.
As @jaltma says, there is massive alpha in being willing to look crazy.
IzwiAI retweeted
300+ GitHub stars for @IzwiAI
1200+ Downloads 🚀
I strongly believe voice AI will be huge in the next few years.
Izwi's goal is to be the default voice AI runtime for private deployments.
github.com/izwi-ai/izwi
Only 4 more needed to reach 300 GitHub stars.
Tell a friend to tell a friend to try Izwi, the Voice AI Runtime.
github.com/izwi-ai/izwi
IzwiAI retweeted
At least we are not buying GitHub stars at @IzwiAI, we need a few more to reach 300. Go to GitHub and give us a ⭐️
github.com/izwi-ai/izwi
IzwiAI retweeted
Being a solo founder in the age of AI is weird, you can do more than ever, but the real bottleneck is still judgment.
AI gives you leverage. It doesn't give you conviction.
Building Izwi has made that painfully clear.
github.com/izwi-ai/izwi
