Pinned Tweet
Listened to 34.2 hours of voice agent calls to build this
Found issues costing $1.2k a pop
Now available for every Voice AI engineers to use
Agustín retweeted
wake up
hit the gym 💪
go to a hackathon
climb coit tower for some fresh air.⛰️
casually win 2nd place with @agustincto 🏆
life is good.
who is building insurance for AI, my agent promised me that a run was going to cost $2, but it was $8
second on the @manufact track :)
they are from argentina 🇦🇷
These guys built a social network for personal agents
Claim your handle at netsi.io
@miguellalfaro @FedeTemari
ask your muse/instinct to signup to netsi.io and add agustin as a friend
These guys built a social network for personal agents
Claim your handle at netsi.io
@miguellalfaro @FedeTemari
These guys built a social network for personal agents
Claim your handle at netsi.io
@miguellalfaro @FedeTemari
me explaining to @julistack how going to a hackathon this weekend will get us closer to PMF
this is going well under the radar, but the future of voice AI it's going to be with models like this one
Pipecat PhoneLLM Alpha 2 🎉 now in Mini and Max sizes!
Max is a "microthinking" model: one sentence of reasoning keeps multi-step conversations accurate yet still under 200ms TTFAT.
More about it here:
pipecat.ai/labs/phonellm-alp…
Weights here:
huggingface.co/pipecat-ai
ray really cooked with this launch
Introducing Rainbrand, the brand layer for agents.
⇢ ingest brand context from anywhere
⇢ turn it into a machine-readable brand system
⇢ use it across agents and products
We’re building brand infrastructure for a world of agents.
Rainbrand, the brand layer for agents
Rainbrand turns brand context from anywhere into a machine-readable brand system that agents and products can use.
Met Pedro when I judged a hackathon in Brasil, he won solo, one of the most talented engineers I have met
Replying to @GadiBorovich
@pedrostanzani (23, Brazil). He grew up in 3 countries. At Tako, he builds Nara, an AI voice recruiter for some of Brazil’s largest companies. On nights and weekends, he built Brazil’s first electoral MCP (50K+ queries) and a Facemash for Brazilian startups (900K+ votes).
















