Let's. Go.
Most TTS benchmarks are broken.
They assume long, clean, studio audio.
Real users give you 2 seconds, recorded on a phone, with noise.
We at @useCamb_AI built MARS8 to win there.
Anyone following this? Curious for folks doing A/B tests
Super appreciate @akshay_pachaar pointing out the dope work of the @resembleai team.
Hardest part of getting Turbo unlocked is infra and I’m proud @TeamHathora sorted this before Christmas so you can try it this week:
models.hathora.dev/model/res…
Which of our 14 regions are you closest to? I’d love to get that hit ASAP to get you the lowest latency possible.
We would not complain if you got us past the Following/Follower ratio as a Christmas present.
Come through for:
- Multi-Modal AI Inference
- Our globally distributed network
- Dank memes (?)
- Niche gaming references
See you soon!
Team Hathora retweeted
Announcing our latest partnership with @TeamHathora!
Hathora is a next-generation model marketplace and inference platform, giving developers unified access to cutting-edge ASR, TTS, STS, and LLM models across 14+ global regions, without managing infrastructure or multiple vendors.
Hathora’s architecture is built for real-time AI, routing every request to the closest available model instance for consistently low-latency performance.
We’re excited to bring Rime’s Arcana model into the Hathora marketplace, making it easier than ever for developers to add our expressive, converting, and enterprise-grade voices to their agentic and conversational AI applications.
Get up and running in minutes with Arcana on Hathora!
Tonight’s AI Lightning Talks at the @shinecapital office with talks from
Zack @AssemblyAI
@andre_slav03 @TeamHathora
And Kent from @Medal_TV and General Intuition.
We lit up the minds of NYC AI curious.
Come back in 2026!
We asked ~200 builders at yesterdays hackathon a simple question:
Which Real-time AI models do you actually want to use?
Some clear winners. Some surprises. A few “wait, really?” moments.
What stands out to you from this chart?
And which models should we add next?
More marketplace insights dropping soon. 📈
Team Hathora retweeted
You might be shocked, but I didn't learn to code in Rabbinical School - but with @base44 i made my first voice-enabled app.
vocal-list-1d7a1e75.base44.a…
Pardon the crappy name and limited utility - but I swapped out the non-working ASR model for @nvidia's parakeet and I pulled out the standard Base44 LLM with @Alibaba_Qwen 3 30b, and this thing works! (Ofc, hosted on @TeamHathora)
I did do this with @Lovable first, but needed help so I can't count that as a totally solo project. Stand back, I'm a vibe-coder now.
Only three companies offer Speech-to-Speech APIs today: @OpenAI (GPT Realtime), @hume_ai (EVI), and now @TeamHathora with Qwen3-Omni.
That's right, we've shipped Qwen3-Omni to production, no waitlist, no queue.
Test it live: models.hathora.dev/model/qwe…
Open-source S2S + real-time inference + multi-region infra.
Let me know what you think.
Bringing back past family members with AI?
A regular Sunday.
I tried my first #voiceAI project on @TeamHathora and recorded it on @loom.
Used the @chatterbox model. I'm going to try it with @ElevenLabsDevs next. Maybe @thorwebdev will help me if I can't figure it out.
I'd love your feedback!
loom.com/share/5c7e3feb68be4…
Team Hathora retweeted
Hello to everyone in the small multiverse of Voice AI. I am Abhishek. I have been building @speechtonote with my agency folks @teamcodesign for the last couple of years.
It's a nice little handy-dandy Micro SaaS application which allows you to play with Voice AI.
Team Hathora retweeted
First demo posted in the Voice AI community on @X.
What's your latest voice AI demo, folks?
Ok, who else is in?
Voice AI roll call:
- @rimelabs - looking at you @lilyclifford and @patrickscoleman
- @AssemblyAI - your team should be here
- @livekit - you're welcome here too
- 🐈 @pipecat_ai - y'all know what's up
Who we missing?
Inspired by the Running Community, the amazing @kwindla meet up in SF, my guy @thorwebdev being the king of @ElevenLabsDevs, and hanging with @jundesai from @cartesia, I realized there is no #VoiceAI community on X so figured to start it and see what happens.
There is a lot of "competition" in this space, but I feel like there are billions of people who talk and it won't be winner take call, so let's at least hang out while it's still fun!
Let's hit 10 people here before the weekend!
teapawt.org/i/communities/19893058…
With yesterday's @ProductHunt launch, we have a leaderboard of our own!
Seems like @nvidia's parakeet 🐦⬛and @chatterbox took the cake.
What models do you want to see next?
Unreal to see us at 170+ upvotes on @ProductHunt - the community is undefeated.
Can we hit 200+ in upvotes by the end of the day? That would be amazing.
Tag a friend who you think would like to use an open source voice model in 14+ regions today!
Team Hathora retweeted
We're about halfway through our launch and are holding solid at #7 on @ProductHunt - if you're a PH person and can give us some feedback that would be amazing.
producthunt.com/products/hat…
Thanks to @benln, @katedeyneka. and @hasantoxr for being some early and generous backers!
