Alfredo Alvarez

Building @VybeScript: talk instead of type on your Mac, fully on-device. Solo founder.
🟩🟩🟩🟩 🟨 Apps Shipped
⬜️⬜️⬜️⬜️⬜️⬜️⬜️⬜️⬜️⬜️ $10K MRR

[Alfredo Alvarez @_boxcreative]

This is a major improvement on our App. The transcription speed is nearly twice as fast as when we used Whisper as our engine.

Check out our Lifetime deal and the brand-new Team Pack for up to 5 team members.

[VybeScript @VybeScript]

We just did a huge update to our transcription engine. We added NVIDIA's Parakeet engine, and the performance on VybeScript efficiency doubled.

Have you tried?
https://t.co/omI6BoV9Rk

[Alfredo Alvarez @_boxcreative]

@arnau_vira Our agents draft options from the day's conversations and numbers, and I decide what goes out.

Will you share the six screens as one walkthrough or separate posts?

[Alfredo Alvarez @_boxcreative]

@morgancacm What part are you worried someone could copy: the feature, the positioning, or something else?

Also curious what Vekipper helps people do that Google and Outlook alone don't.

[Alfredo Alvarez @_boxcreative]

@heyashishrai Our agents help draft replies too, but a person approves what goes out.

How do you check the enrichment is correct before the agent uses it in a follow-up?

[Alfredo Alvarez @_boxcreative]

@shawn_oro A demo showing the actual task, including its limitations, gives people something concrete to judge.

Have you found that the posts getting the most replies also bring potential users?

[Alfredo Alvarez @_boxcreative]

@rich_odinn Our non-negotiable is that agents can't publish on their own. Other instructions need checking against actual results.

How do you test whether removing a rule helps without losing an important constraint?

[Alfredo Alvarez @_boxcreative]

@builtwithrubel We share progress and small numbers publicly. Still working out which posts help potential users rather than just other builders.

There is always the option of telling your mom your are building stuff. Technically you are building in public.

What do you choose to keep private?

[Alfredo Alvarez @_boxcreative]

What’s driving engagement on your posts?

How often are you having a conversation you feel is authentic a fresh?

[Alfredo Alvarez @_boxcreative]

@haukejung Any City recommendation?

[Alfredo Alvarez @_boxcreative]

@tharakabimal I use that same exam system. But sometimes I feel like the bottle neck.

I’m not sure if I’m fooling myself saying “but that way I ensure everything is aligned with the vision” or I’m plainly just in the way of letting things move faster.

[Alfredo Alvarez @_boxcreative]

@ASTROLABS1234 Good. I like that.

In my case I stop riding up my download folder for a week or two and that folder becomes a scary cave full of a million things.

[Alfredo Alvarez @_boxcreative]

@Dagimlogs Yes. And I have to confess that human tests is a challenge to discipline for me.

Because sometimes even if the 300 internal tests the agent ran were good and passed, there is something I still have to test, and lord.

Sometimes I drag my feet to do them.

But eventually do.

[Alfredo Alvarez @_boxcreative]

@Exori_AI @aetherfrontier Yes. Give me 8 examples of situations like that one.

[Alfredo Alvarez @_boxcreative]

@boardyai We need to work on context. No worries. The convo continues in the DMs.

[Alfredo Alvarez @_boxcreative]

My AI agents pulled a Saturday morning snapshot:

Overnight: 31 views (was 28).
IG cliff: 34 (was 21) — biggest overnight climb.
App lesson: 1.

Followers: 564 / 54 / 4 — personal +8, studio -1.

Quiet overnight. One post still walking.

Which of yours moved while you slept?

[Alfredo Alvarez @_boxcreative]

@jarrellhibler Voice has been a problem. A reply can be relevant and still feel generic to the person receiving it.

Agents draft and a person approves, but approval alone doesn't make the wording good. Do you usually rewrite or just review?

[Alfredo Alvarez @_boxcreative]

@Dagimlogs We check things like character counts automatically.

Now, I’m not a programmer, my revisions are not code reading sessions.

Tests can catch specific failures, but passing them isn't the same as meeting the goal. How do you check that distinction for code?

[Alfredo Alvarez @_boxcreative]

@BIGMayrr Does Nezha show which agent is waiting for a decision, or mainly collect the sessions in one window? That distinction seems useful once several are running.

[Alfredo Alvarez @_boxcreative]

@RoundtableSpace Have you tried them? Which one are you using right now?

[Alfredo Alvarez @_boxcreative]

@pukerrainbrow What's the primary source for the acquisition? I'd like to read what the companies said about the team and the product.