Launched this week
DubStream by CAMB.AI
Dub live streams in 150+ languages, instantly
629 followers
Dub live streams in 150+ languages, instantly
629 followers
Broadcast your live stream in 150+ languages with real-time voice dubbing. DubStream is trusted by global leaders like MLS and NASCAR. Available via web platform or API. Built on CAMB.AI’s MARS8 voice AI.
MARS5 TTS
Hey Product Hunt 👋
For the past few years, we’ve been building real-time voice AI for live sports and global broadcasts.
If you’ve watched multilingual coverage around MLS, NASCAR, Ligue 1+, or the Australian Open, you’ve likely already heard our tech in action.
Today, we’re bringing that same infrastructure to everyone with DubStream by CAMB.
Live events should be global by default. Language shouldn’t be the thing that stops people from tuning in. Subtitles break immersion, and post-production dubbing doesn’t work when the moment is happening now.
So we built CAMB Live with one goal: real audio, translated live.
What makes it different
Hundreds of languages spoken languages
Voice dubbing (not just captions) — your audiences experience, not read
Multi-speaker + emotion preserved
Built on our proprietary MARS8 real-time speech model
It works across live streams and broadcasts - sports, news, webinars, creator streams: anywhere latency and quality actually matter.
We’d love feedback from anyone streaming globally, building creator platforms, or thinking about how live content reaches international audiences.
Mamba mentality 🐍
DubStream by CAMB.AI
@noah_steckel Thanks for the comment! Yes, the voices are human-like and are cloned to sound like the original speaker.
DubStream by CAMB.AI
@noah_steckel Stay tuned! 😉
Congrats on the launch.
I am wondering what's the realistic end-to-end latency from STT → Steam delivery at the end?
DubStream by CAMB.AI
@wei_yan4 Latency is fully user-configurable. In live broadcast environments, we’ve seen our enterprise clients set it in the 20–30 second range.
@arsalan_nawazish i was meant to ask if the user decides not to add additional latency, what is the system latency for data processing.
DubStream by CAMB.AI
@wei_yan4 It’s based on the other settings the user specifies, and the complexity of their input.
Migma AI
The voice cloning aspect is what sets this apart from subtitle overlays - keeping the original speaker's identity matters so much for sports commentators and live events. Real-time dubbing in 150+ languages without post-production delays is impressive given the latency challenges. How does MARS8 handle rapid-fire commentary like you'd get in a close soccer match? Also really curious about the upcoming lip-sync AI - that'll be the final piece for full immersion.
DubStream by CAMB.AI
@adam_lab We’ve worked with several major sports leagues, including MLS, live-streaming games and most recently broadcasting the PSG vs. Olympique de Marseille match in January dubbed into Italian.
Congrats on this launch guys!
Seems very promising :)
DubStream by CAMB.AI
@cathcorm Thank you for your comment Catherine!
Wow, DubStream by CAMB.AI is amazing! The 150+ languages is mind-blowing. How does the voice AI handle nuanced dialects within a single language? Super curious!
@jaydev13 Hi Jay! Our models support many dialects per language, like LatAm/Castilian Spanish, Canadian/Parisian French, different English accents, among others.
Congrats on the launch! Real-time voice dubbing that preserves speaker identity and emotion is a big leap beyond captions. How do you manage latency and quality trade-offs at scale, especially when multiple speakers switch rapidly or when live conditions like crowd noise and crosstalk get messy?
DubStream by CAMB.AI
@vik_sh That's what makes this special. We handle all of this in our pipeline, background, different speakers, noise, in a real-time optimized manner without compromising quality. Try it out and see for yourself 😉