uRun

This account is verified. Learn more

@urunml

Real-time AI model APIs that run together, for voice agents, live video, and avatars.

San Francisco, CA urun.sh Joined March 2026

3 Following 33 Followers 47 Posts

Latest posts 5

uRun @urunml 1d

Huge congrats to our friend @rissa_cao and the @FishAudio team on Live 3!

A voice model that can say “mm-hmm” while listening, but knows when to stop at “wait, actually...” is a big step towards more natural, real-time conversations.

Go give it a try! 🚀

x.com/rissa_cao/stat…

Rissa Cao Verified account @rissa_cao 1d

Introducing @FishAudio Live 3: a full-duplex voice model that listens, says "mm-hmm," and knows when to shut up.

Enterprise? Book a demo for your free custom agent: s.fish.audio/vxhgs6

Reply with your use case - we'll build it free for the first 1,000 people.

View quoted post on X

uRun @urunml 5d

Model APIs are live on uRun.

In the demo, we plug Qwen 3.8 27B into OpenCode and chat with it over industry-standard chat completions. Working internally, rolling out worldwide.

Our waitlist gets access first: urun.sh/contact

More: urun.sh/blog/model-api…

0:36

uRun @urunml Sep 4

almost everything in ai is still optimized for stateless request and response. that falls apart the second you want video or audio a user can interrupt and steer.

@keeganmccallum3 on tank talks, on why the state belongs on the server 👇️

x.com/mattybcohen/st…

Matt Cohen Verified account @mattybcohen Sep 4

How do you scale 500 to 9,000 H100s in six hours!!

That's what it took to keep @DreamLabLA Dream Machine standing when it went viral. Half a million videos in 12 hours.

@keeganmccallum3 is now building @urunml for the interactive AI wave.

New Tank Talks [links below]

View quoted post on X

uRun @urunml Aug 21

$50 buys a day of continuously generated video. plenty of engineers spend that on coding tokens in an hour.

our ceo @keeganmccallum3's @aiDotEngineer talk on what that changes is live. thanks @swyx + team 🙏

new design partners: our dms open 👋

youtu.be/Xln-On3syJk

www.youtube.com Generative Video at the Speed of Light — Keegan McCallum, uRun Ten dollars now buys roughly three hours of continuously generated ...

uRun @urunml Aug 6

fast mode costs 2.5x and it's still waitlisted. the market is paying multiples for presence, not batch jobs.

@keeganmccallum3 on why we make that the default 👇

Keegan McCallum Verified account @keeganmccallum3 Aug 6

OpenAI and Anthropic charge ~2.5x for fast mode. Even then, OpenAI downgrades priority traffic and Anthropic runs a waitlist.

That's the tell: a continuous, responsive session is worth multiples of a job you fire off and check on later. At @urunml build for that by default.
View quoted post on X