Kinward

a privacy first, easy-install local AI for families, businesses and people who want their data back.

[Kinward @trykinward]

As we predicted, the tide is shifting to local LLM solutions and the benefit of hosting your own AI model is becoming easier.

Get a head start at https://t.co/gqcrtVOAXk

[Kinward @trykinward]

@cursor_ai @elonmusk @SpaceX Congrats!

[Kinward @trykinward]

@MichaelDell We're building Kinward, the family layer on top of local AI. Per-kid profiles + guardrails, memory of your whole household, governance without surveillance. Model-agnostic, runs 100% on your hardware and never the cloud.

Free alpha now, Mac + Windows. 🛡️

[Kinward @trykinward]

Kinward is live. Navigate the setup of a local LLM with ease.

Your family's AI — on your own computer. No cloud. No accounts. No telemetry.

Everything you share with it stays inside your house. That's it.

Free alpha · Mac + Windows · https://t.co/3Oe4P0Sera

[Kinward @trykinward]

“If you want AI for your family, you shouldn't have to give your family to a tech company."

Built with that philosophy. Free alpha. Open source. https://t.co/9Hh5wDjLdA

[Kinward @trykinward]

@LottoLabs Tailscale will be packed on our install for non-dev families. A number of folks were asking about remote connection while protecting their data. This week overall has been huge for local LLM developments.

[Kinward @trykinward]

Kinward + Gemma 4 runs on a mid-range PC. New Alpha update live on our repo:

kinward-ai/kinward: kinward ai

If you've been waiting to setup a local LLM for your family but unsure of how to do it, we're here for you!

[Google @Google]

We just released Gemma 4 — our most intelligent open models to date.

Built from the same world-class research as Gemini 3, Gemma 4 brings breakthrough intelligence directly to your own hardware for advanced reasoning and agentic workflows.

Released under a commercially permissive Apache 2.0 license so anyone can build powerful AI tools. 🧵↓

[Kinward @trykinward]

New COPPA rule goes into effect on 4.22. In 20 days, AI orgs face stricter rules.

Our approach: don't collect it in the first place. Run local. Keep everything on your own hardware. No children's data to regulate if it never leaves your home.

https://t.co/9Hh5wDjLdA

[Kinward @trykinward]

@YueKuratsu I remember using these for about a year.

[Kinward @trykinward]

@davepl1968 Amazing.

[Kinward @trykinward]

@DeFi_Hanzo Literally news everyday that favors local LLM optimization. Exciting times...

[Kinward @trykinward]

Google's TurboQuant: 6x memory compression, zero accuracy loss, no retraining, already being ported to llama.cpp and MLX.

If you're 🔨for local AI on consumer hardware, your users just got 6x more context instantly.

Households are getting more power...

[Kinward @trykinward]

@DeFi_Hanzo This is a big step for local LLM performance. Excited to how this is integrated into all of the local models very soon.

[Kinward @trykinward]

@LottoLabs And your data is yours!

[Kinward @trykinward]

My friend shared his new Mac mini installing the https://t.co/gqcrtVOAXk alpha! The Kinward army is growing!

[Kinward @trykinward]

@thestreamingdev very cool!

[Kinward @trykinward]

@thestreamingdev @TheVolun Indeed. Not everyone needs large-B models...

[Kinward @trykinward]

@ekimroyrp @threejs so dang beautiful

[Kinward @trykinward]

@ShimazuSystems Very true, and those within dev circles can figure most of this out.

I do think there is something to be said for a packaged wizard to help non-devs get a local solution up and running.

[Kinward @trykinward]

And because LiteLLM is a transitive dependency in tons of AI agent frameworks, many teams got hit without ever running `pip install litellm` themselves.

I'm building an AI platform for households that runs entirely on local hardware via Ollama. No cloud calls, no API keys.

[Kinward @trykinward]

The LiteLLM attack is a nightmare for cloud-dependent AI stacks:

1. Trivy compromised (March 19)
2. Checkmarx KICS compromised (March 23)
3. LiteLLM compromised (March 24)
4. SSH keys, cloud creds, crypto wallets, .env files — all harvested