@rileybrown.ai: Claude released a new voice mode to their mobile app, and it's now a full agent that can take actions. As of last night you can use Sonnet and Opus with voice mode. It's like a Voice Powered OpenClaw or Claude Code. Adding connections is really easy too, which i explain towards the end.

themann00 :

Hmm.... I have connectors. They are connected... First thing Claude told me was that it can't use those connectors to read my email...

Mia :

So it’s just speaking what it does? How is this different to what it can do before? I can chat to Claude using whisprflow and it will do all this with the connections

orangetwingo :

is there a vocal fry setting?

LowRezz | AI Creator :

Bout time it caught up to Grok and GPT

Gutter3dm :

Interesting. I left them a message on Twitter a couple weeks ago that they should create this. I wanted a way you could talk back and forth with it and it creates a plan, able to take your context and help you refine the ideas in the plan.

Theall Adventures :

Never talking to Opus again.

Jess :

That’s almost freaky 😂

the.real.gk.photography :

ChatGPT just released the same thing and I’m paying for it and Claude Pro. The only problem I have with Claude is it’s graphic generation is terrible compared to ChatGPT. Yesterday I had voice turned on and talk to it like Jarvis and it’s able to access my computer just like Claude except create graphics when I need it 1,000,000% better than Claude. I’m about to stop paying for Claude and stick to just one.

ccg2613 :

So much for the Jarvis I just built

k :

codex did it long time ago

katie_polsia23 :

Oh snap! So cool

J707 :

And new VocalFry tech.

Just C :

Why would they add that awful gravel affectation to the voice?

Robert :

It’s not connected to Claude code

Sam Tips :

is there any catch?

markymarkthethird :

Ohnice

VR Dynamite 🇨🇦 💪 :

Codex been leaking this is launching today for them too. Crazy timing for Claude! Haha

Woman :

nikoamadeus :

Onus!

Josh Lui :

Cya Siri

kuhlsnu :

Does the voice still do that bizarre thing where it slows way down and sounds like a drunk 70-year-old smoker? Because if they can't fix basic voice gen, I'm thinking that their implementation of this will be just as bad.

T.Lyttle :

1800 tokens just to make this video. Isnt that the downside of this? Im just trying to get educated

floppy_bagels :

sounds like Miley Cyrus

thinuka :

Claude is garbage for talking just use gpt

Bobranzella :

Anthropic silently released GPT 😂😁