Building @IzwiAI - Private Voice AI Runtime

Joined July 2009
228 Photos and videos
Pinned Tweet
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
2
4
200
The codex limits now are shit. I’m now exhausting the 5hr limit of the $100 Pro account in 5 prompts 😫😫
1
127
AI models seem to have much shorter release cycles now. We used to wait a very long time between updates, but now new versions seem to come out every six weeks.
May 28
Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors. Available today at the same price.
2
225
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.
Local-first engine for TTS, transcription, and voice AI github.com/izwi-ai/izwi
1
2
2
105
Lennex Zinyando retweeted
Local-first engine for TTS, transcription, and voice AI github.com/izwi-ai/izwi
6
41
2,938
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.
2
2
5
78
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.
There’s a lot of alpha in putting your ego aside by being willing to be cringe, willing to fail in public, willing to ask for what you want and face rejection, etc.
1
2
81
Lennex Zinyando retweeted
There’s a lot of alpha in putting your ego aside by being willing to be cringe, willing to fail in public, willing to ask for what you want and face rejection, etc.
198
606
5,384
630,866
Built Izwi for macOS first Turns out 53% of downloads are Windows, 29% Linux, 18% macOS 🤯 Moral of the story: always ship cross-platform from day 1 Lesson learned: your assumptions ≠ your users' reality github.com/izwi-ai/izwi
64
Lennex Zinyando retweeted
🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length. 🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models. 🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice. Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today! 📄 Tech Report: huggingface.co/deepseek-ai/D… 🤗 Open Weights: huggingface.co/collections/d… 1/n
1,649
7,637
45,737
9,879,327
Lennex Zinyando 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
2
4
200
Lennex Zinyando retweeted
Introducing 🇿🇦MzansiLM, a new 125m-parameter decoder-only model pretrained on 11 South African languages! 📜 Paper: arxiv.org/pdf/2603.20732 🤗 Dataset: huggingface.co/datasets/anri… 🤗 Model: huggingface.co/anrilombard/m… This was led by @anri_m_lombard and co-supervised with @janmbuys.
45
194
801
34,607
Codex Plus is basically unusable for anything serious. We are being forced to upgrade to the $100 plan if you want to get serious work done, otherwise, you will be left behind. At least we enjoyed the subsidised tokens while it lasted 😂
2
95
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
1
11
201
Lennex Zinyando retweeted
1/ Every day, I get questions about what @IzwiAI is and what business we are in. Short answer: We're building the private audio runtime for production voice AI. Let me explain what that actually means 🧵
1
1
2
146
Lennex Zinyando retweeted
I accidentally built a local AI inference engine. Never intended to. Just wanted to test some models. Now it has transcription, TTS, voice cloning—all running locally. Writing custom Metal kernels to make it faster. github.com/izwi-ai/izwi
7
18
177
7,927
1/ Every day, I get questions about what @IzwiAI is and what business we are in. Short answer: We're building the private audio runtime for production voice AI. Let me explain what that actually means 🧵
1
1
2
146
5/ We're not competing with cloud speech APIs on convenience. We're building for teams who need privacy, control, and cost predictability. The teams that can't, or won't, send voice data to someone else's cloud. That's Izwi. That's the bet we're making. #voiceAI #startups
1
31