300+ GitHub stars for Izwi
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
At least we are not buying GitHub stars at Izwi, we need a few more to reach 300. Go to GitHub and give us a ⭐️
github.com/izwi-ai/izwi
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
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 #AIinfrastructure #startup
4/ Izwi gives you one unified runtime that runs entirely on your own infrastructure, on-device, in your VPC, behind your firewall.
OpenAI-compatible API. No audio leaves your environment. No usage-based billing that explodes at scale.
Just production-ready voice infrastructure you control.
3/ Neither works when you're a healthcare company that can't send patient data to the cloud. Or a fintech with strict data residency rules. Or any team needing sub-100ms latency.
The market is telling us: private deployment is no longer a niche feature. It's a standard buying requirement.
2/ If you're building voice AI today, you have two options:
1. Send audio to the cloud (Google, AWS, OpenAI) - easy to start, but expensive at scale + data privacy headaches
2. Stitch together open-source tools - each with different APIs, barely production-ready
Neither works for many teams.
1/ Every day I get questions about what Izwi 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 🧵
How do you build features people want when nobody opts into telemetry?
Open source founders know the trap: ship, celebrate, nada. Just vibes and prayer.
Izwi launched with optional telemetry. <1% opted in.
Now we're flying blind.
Startup problems.
github.com/izwi-ai/izwi
Fireworks AI shutting down their audio APIs is a reality check.
Given 30 days' notice, you realize the brutal truth: You’re not building a product. You’re renting one.
Portable audio inference. Run it anywhere, cloud or on the edge. Own your stack.
github.com/izwi-ai/izwi
I think OpenClaw and Hermes Agent are incredible, 350K+ and 28K+ stars are incredible. But they've been stuck in text.
Enter Izwi: on-device voice AI that runs locally.
Your agents, now with ears and a mouth. 🦞🗣️
github.com/izwi-ai/izwi
ElevenLabs biggest flaw: you're locked into their cloud.
Their servers. You pay per character. You need API keys.
I built Izwi AI.
Open-source voice AI that runs 100% on your machine.
No API keys. No cloud. No subscription.
276 ⭐ on GitHub.
github.com/izwi-ai/izwi
🚀 Just shipped v0.1.0-beta-11, the biggest UI update in this beta cycle.
Redesigned UI + new features that make Izwi feel like a real product.
Hot take: "Open source ElevenLabs"
Try it and tell me I'm wrong: github.com/izwi-ai/izwi...
⭐ If you believe in open source voice AI
ChatGPT is now the #1 traffic source for the Izwi website.
400 sessions this month. More than Google. More than Twitter.
The AI search era is here.
github.com/izwi-ai/izwi
ElevenLabs recently wrote about open-source TTS for conversational AI.
Soon, Izwi will be THE way to run open source ASR, TTS & Chat models.
GitHub: github.com/izwi-ai/izwi
Question I keep asking myself: Are we building infrastructure or building a product?
Izwi is the Ollama of Audio AI or the open source ElevenLabs?
Still figuring out which lane to pick.
Hot take: this is exactly when a cofounder earns their keep
github.com/izwi-ai/izwi
Izwi just shipped Studio, a workspace for long-form TTS projects.
Paste text, split into segments, queue renders, export audio. It's beta and rough in places, but it works.
Try it and tell us what's broken.
github.com/izwi-ai/izwi...
The product decisions I didn't expect: they're all local first problems.
Speed depends on hardware. Bigger models are smarter, but don't run on weaker machines. Features behave differently across models.
Making it all feel cohesive is the job.
github.com/izwi-ai/izwi
Getting paid users is the next step and it’s probably going to be hard. I like that the idea, problem and solution have been validated. Building a business is next.
Day 1: "You need a co-founder."
Day 60: still solo, 220 GitHub stars, shipping voice AI that runs on your machine
Turns out the math changed.
Building in public. Izwi. Follow along.
github.com/izwi-ai/izwi
AI makes building easy. Knowing what to build? Still, the hard part.
This Saturday: UI polish (I'm not a designer, it shows) + competitor intel.
AI helps you build. It does not help you decide.
Anyway. Back to working on IzwiAI.
github.com/izwi-ai/izwi
I asked Grok to tell me about myself, and it said:
"You're the guy quietly (but very deliberately) building infrastructure so voice AI can actually work offline, cheaply, and, hopefully, soon in languages and contexts that big cloud providers still ignore."
IzwiAI 🚀🔥
200+ stars on GitHub for Izwi.
This project started because I wanted to try Qwen-TTS/Qwen-ASR and couldn't find an easy way.
So I built it myself.
Started by accident. Feels like the right move.
github.com/izwi-ai/izwi
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
"We have vendor assessments." → "Our AI never touches the cloud."
"We have DPIA underway" → "GDPR-compliant by design."
The difference between a 3-month sales cycle and a 12-month compliance review.
izwiai.com/blog/gdpr-by...
Your M&A strategy lives in vendor logs. Your SEC investigation discussion sits on someone else's servers.
That's not a risk. It's a certainty.
On-device transcription keeps your boardroom in the room.
izwiai.com/blog/ai-in-t...
