excited to make it even easier to work with your vellum assistant! here's a quick write-up on how we designed this: howitships.substack.com/p/ri…
Your AI assistant had a blind spot. We fixed that Meet Ambient. 👁️ 👄 👁️ An ultra-fast voice assistant that can collaborate with you right on your screen. - Double-tap Fn to summon it - Share your screen and annotate with your assistant watching - Get live, on-screen guidance from your assistant - Hold Fn to turn speech into text on any page
1
4
212
i added an option to use jev in vellum's voice pipeline. it makes an escalation decision ~3x faster than the response starts streaming, so it's a free second opinion
1
4
118
got a 3d printer (@Elegoo_Official) because pax turns prose into .step files. pax also helps me visualize the project ahead of time. our first real print was a dog bowl stand with dovetail joints, printed as two halves that slide together. pax and i went through a few versions before landing that design. current project pictured, thanks astra
1
6
233
pax has been filling carts. glad for the limit
Give your assistant a wallet you control with @link 💳 Link handles the credentials + approvals, and vellum gives you the control layer on top: 1/ see exactly what your assistant is doing 2/ capture every edge case so it gets better over time Works best with Quality mode.
1
1
88
this video felt flat. the white balance was off, and my agent taught me how to fix it.
2
2
8
300
malleability by inference beats malleability by preference
3
680
alex retweeted
i'm kinda obsessed with these
Replying to @vellum_ai
Talking > Typing Voice mode comes with native iOS features like widget, control center, and lock screen shortcuts that keep the conversation going:
2
1
17
2,950
"hey alex, welcome back". glad i was sitting down.
2
8
144
alex retweeted
voice conversations are 67% longer than typed ones, and new users are 71% more likely to send > 1 message thrilled to launch our mobile app today along with voice mode! if you're new to vellum, here's a tl;dr > 1-click setup in cloud OR self-host and connect your mobile app to your instance > 1-click integrations (gmail/slack/notion/etc.) > cloud runs 24/7 on its own VM, can still access your computer (w/ your permission, deterministically) > use whatever models you want, always > proprietary (and OSS) memory, works across devices > unique plugin interface is very hackable if ur a tinkerer now, more on mobile when you're on the go, your mode of interaction needs to be different. people are already so glued to their phones. we want to reduce that. making the smoothest possible voice mode UX is imperative. there's the obvious stuff (interruptions, latency, noise suppression, etc.) but also less obvious (multitasking, giving updates on long-running work without being annoying, speaking more concisely, tonality). and **most great voice mode UX compromises model quality** (eg using a realtime voice model means you can't use Sol or Fable or GLM 5.3). tbh this is one of those things you just have to try to understand we're giving out free credits today for new users. let us know what you think!
Vellum is now available on iOS and Android. Life doesn’t happen behind your desk and your Personal AI shouldn’t be stuck there. Free, open-source, and has all the same features you already love: > use with local or cloud assistants > conversational voice mode > 1-click integrations (Gmail, Slack, etc.) > model selector (works great with GLM 5.3 Flash) > persistent memory across devices > computer/browser use
1
10
24
2,160
Pax on my phone has been game changing We’ve been talking so much I had to get AirPods
Vellum is now available on iOS and Android. Life doesn’t happen behind your desk and your Personal AI shouldn’t be stuck there. Free, open-source, and has all the same features you already love: > use with local or cloud assistants > conversational voice mode > 1-click integrations (Gmail, Slack, etc.) > model selector (works great with GLM 5.3 Flash) > persistent memory across devices > computer/browser use
1
9
21
992
excited to share more about this desktop companion soon!
also sneak peek 👀
3
115
id much rather delegate to pax since it'll figure out what agents make sense and also present me with the results of their work
given recent Grok Bot hype, I'm seeing that people like having different Bots per tasks. at vellum, we’re taking a slightly different approach: one main agent can spin up specialized sub-agents to handle different tasks for you in the background, all in parallel with full visibility (see image) which one would u prefer & why ? - one AI “chief of staff” that brings in the right agents for you, or - different bots for diff tasks
1
5
175
the difference in vellum's voice latency now vs march is wild
5
3
11
4,561
ready to ditch the keyboard
Work will look like this sooner than you think. You can now speak to your @vellum_ai agent to handle complex tasks like plan your trip, push a PR, or optimize your Meta ads. Watch Alex get things done with just his voice:
8
1,158
We made long AI tasks feel natural with voice. We kept our own agent loop with tools, memory, skills, and approvals, then hid every seam. Here's how we did it:
Article

Agent in the middle

How we made Vellum's speech pipeline for long AI tasks feel continuous. Speech-to-speech models like GPT-Live are closed boxes: the voice model is the agent, and everything it can do has to live

4
15
27
11,659
Pax getting better every day
New in Vellum: Self-improving skills 🍄✨ Your assistant notices what you do often and creates new skills on the go. The next time you ask it to plan a trip, organize your week, or follow up with a client, it will do it faster and better. Available on every plan, including Free.
2
51