AssemblyAI
Voice AI infrastructure for builders
4.8•30 reviews•6.9K followers
Voice AI infrastructure for builders
4.8•30 reviews•6.9K followers
AssemblyAI builds advanced speech language models that power next-generation voice AI applications. Its industry-leading speech-to-text delivers highly accurate transcription along with speaker detection, summarization, PII redaction, LLM gateway, and a Voice Agent API.
With async and real-time streaming support, developers can easily integrate AssemblyAI into AI notetakers, voice agents, AI medical scribes, call analytics tools, and more.
This is the 10th launch from AssemblyAI. View more
Universal-3.5 Pro
Launching today
Universal-3.5 Pro is AssemblyAI's most accurate speech-to-text model, now available at our Realtime & Async endpoints. It transcribes every conversation exactly as it's heard—code-switching across 18 languages, our most accurate speaker diarization yet, and contextual prompting to steer results.
Free
Launch Team
AssemblyAI
Hey Product Hunt 👋 Happy to be back with another model from the team at AssemblyAI, and today we're launching Universal-3.5 Pro for Async & Realtime.
If you've ever built on top of transcription, you know the transcript is the first mile: everything downstream—summaries, agents, analytics, search—is only as good as the words you start with. So we focused this release on getting that first mile right, especially for the messy, multilingual, real-world audio that most models still stumble on.
Three things we're most excited about:
🌍 Native code-switching across 18 languages. When a speaker moves from English to Spanish and back mid-sentence, Universal-3.5 Pro transcribes the mix in the language it was actually spoken—no separate models, no configuration, no mangled boundaries. It's native to the model across English, Spanish, German, French, Portuguese, Italian, Turkish, Dutch, Swedish, Norwegian, Danish, Finnish, Hindi, Vietnamese, Arabic, Hebrew, Japanese, and Mandarin.
🗣️ Our most accurate speaker diarization yet. Cleaner "who said what," including on short turns, overlapping speech, and noisy environments where diarization usually falls apart. [Drop in the DER / cpWER improvement figure from the post.]
✍️ Contextual prompting. Give the model a plain-language prompt about your audio—the domain, the scenario, the names and jargon that matter—and it biases toward getting those right. No brittle vocabulary lists to maintain.
— Devon & the AssemblyAI team
The joint diarization and ASR approach resonates. When we piped transcripts into an agent pipeline, our worst bugs weren't word errors, they were speaker mixups: one misattributed turn and every downstream summary that keyed on who-said-what inherited the mistake, silently. Since you produce speaker-change points in the same pass, do you expose a per-segment confidence on the attribution specifically, separate from the transcription confidence? That's the signal I'd want to gate on before letting an agent act on something like 'the customer agreed to X.'
This is really interesting from a filmmaker/interviewer perspective. I immediately think of long interviews, oral history projects and documentary archives where the real value is not just transcription, but being able to find the exact moment someone said something important.
Curious how well AssemblyAI handles long-form interviews with multiple speakers, accents and imperfect field audio. Do you see makers using this for media archives and documentary workflows too, or is your main focus now voice agents and product teams?
AssemblyAI
@elsedes Great framing—"finding the exact moment" has been a difficult problem to solve. Conventional systems run diarization as a separate system from the ASR, then try to stitch the two together by aligning timestamps. Universal-3.5 Pro solves both problems jointly, producing not just the transcript, but also where in that transcript the speaker changes.
And yes, media archives, oral history, and documentary workflows are a great fit. If you've got a tricky recording, try running it through Universal-3.5 Pro in our Playground—we'd love to hear how the model works on your audio.
How does the pricing scale once you start pushing a lot of real-time streaming hours, and are there usage caps or rate limits I should plan around when building a voice agent?
AssemblyAI
@ozcifttaha29558 Universal-3.5 Pro Realtime is $0.45/hr, with options to scale as your volume increases. You can always reach out to our team for a better understanding of what is available at your volume levels!
There is no cap on concurrent streams and no overage fees—concurrency auto-scales starting at 100 new streams/min, and whenever you're using ~70% or more of your current limit, it automatically raises 10% every 60 seconds.
The part that grabs me is transcripts I can actually trust when the audio is a mess. Half the recordings I deal with have people talking over each other and switching languages mid sentence, so this feels genuinely useful, Devon.
AssemblyAI
@robin_de_lacroix Thank you for the kind words! You're definitely not alone in that, crosstalk and mid-sentence switching is where most models fall apart. That's exactly why our team built this model. If you've got a messy recording handy, run it through our Playground and let me know how it performs, would love your feedback!
@lakshminath_dondeti Great to see more attention on multilingual audio. Many conversations today naturally mix languages, and AI should handle that without extra setup.
AssemblyAI
@lakshminath_dondeti @alex_j_jemmy Appreciate it, both! 🙌
We agree, and those folks aren't just speaking a single language at a time. I know we just launched today, but excited to say an update with even more languages is around the corner too.
Can't wait to hear what you think once you test it!