Watching VertoX improve week by week is exciting
by•
One of the things I enjoy most while building VertoX is seeing the technology improve week by week.
Patrick Lumban Tobing, our AI/ML Engineer, recently integrated Qwen3-TTS Streaming into our Streaming Speech Translation framework, and honestly, it’s exciting to watch everything slowly come together.
Lower latency.
More natural voice output.
Better real-time multilingual communication.
There’s still a lot to improve, but seeing the system evolve like this is a great feeling.
It also makes me appreciate how powerful the open-source AI ecosystem has become. Huge respect to NVIDIA, Google, Qwen, and all the teams pushing this space forward.
A lot more coming soon for VertoX 🚀
53 views
Replies
What is your latency of your requests of Qwen3-TTS Streaming service?
@lyshen Right now, we’re still actively optimizing the pipeline, but the goal is sub-second to ~1.5s realtime response depending on the language pair, context length, and hardware configuration.
We’re continuously working to reduce the latency even further and optimize the system. As you know, every language behaves differently; for some languages, the context becomes clear immediately, while for others, the meaning is only fully understood after more of the sentence/conversation is completed because of the grammar structure. So a big part of our work is making sure translation quality stays high without increasing latency too much during natural human conversation.
@nemos Thank you for your sharing. The pipeline is really important. Good luck.
@lyshen Thank you, you too.
@lyshen By the way, are you building a Voice Agent system? I noticed VoiceClaw on your profile.
Would be great to connect. I really enjoy talking with other founders, especially people building in the voice/AI space. I’m also planning to open a Discord community soon and would be happy to have you there as well.
Honestly a beautiful feeling.
Seeing the progress of my product from the side tool it was months ago into a full-blown SaaS application is beautiful.
@dumebio Honestly, I completely relate to that. Watching something grow from an early idea or small tool into a real product is one of the best feelings while building, especially when you start seeing real users and progress little by little.
@nemos Indeed!
Nice this feels like real “infra getting better week by week” progress, not just product hype. Lower latency + more natural TTS is exactly what makes real-time translation start feeling usable instead of experimental🚀
@lamont__justinn Really appreciate that, thank you! That’s exactly what we’re trying to focus on, real infrastructure progress, not just hype. I’m also planning to share more updates today on LinkedIn and here as well, especially around our open-source work and the traction some of our models/projects got in the first few days after release. Still a long way to go, but it’s exciting to see everything improving step by step 🚀