Pinned Tweet
Speech providers mistranscribe around 1 in 4 medical terms. ❌
That's why we built Oak.
Introducing our new multilingual medical speech-to-text model. 🌳
Building a voice agent or speech application? Start with working code. 👇
@Speechmatics Academy is our open-source collection of runnable examples, from multilingual transcription and Agent STT to integrations with popular voice AI frameworks.
2
111
Find an example for your stack, clone the repo and follow the README to get started.
Python, JavaScript, voice agents, multilingual transcription and more.
Explore the code:
github.com/speechmatics/spee…
34
Speechmatics retweeted
Replying to @Speechmatics
Cool! Wha benchmark do you use to evaluate accuracy for medical-grade voice AI?
The Great Escape (from the cloud). 🌊
On his way to Alcatraz, James put @Speechmatics
On-Device to the test against the noise and dodgy wifi of San Francisco Bay.
See how he fared and request early access to our On Device here: speechmatics.com/speech-to-t…
2
210,518
Speechmatics retweeted
YOUR VOICE AGENT IS ONLY AS SMART AS WHAT IT HEARS.
Brilliant LLM. Great prompt. All the right tools.
then one misheard 'no' can still derail everything!
that’s what makes @Speechmatics' new 'Linden 1' so awesome.
It powers Agent STT, a speech-to-text API built specifically for voice agents.
.. and the demo below is seriously cool too.
You watch words appear live, settle into complete speech segments, and align with turn boundaries for LLM handoff.
Separate speech-activity signals make the interface react as you talk.
You see exactly what your app receives and when the model can take over.
then come the numbers:
→ 369ms median time to final segment
→ 0.96% pooled semantic WER
Both listed in Pipecat’s current independent benchmark.
.. and there’s more!
> 55+ languages
> speaker attribution
> up to 1,000 custom vocabulary terms for your product names and specialist jargon
Watch @Speechmatics' demo below ↓
Day 1 at #HLTH2026! 🎙️
We're showing why @Speechmatics is the #1 choice for medical-grade Voice AI.
Stop by Booth V-4025 to put Oak Real-Time to the test across 55+ languages, meet the team, and grab a lucky bag. 🎁
That's a wrap on #VapiCon! 🎤
From rapping to impressions, you put our On-Device voice AI to the test, running completely offline on @nvidia Jetson.
Thanks to the @Vapi_AI team for another brilliant year, and to everyone who stopped by!
Curious about On-Device speech recognition? Let's talk.
speechmatics.com/speech-to-t…
Speechmatics retweeted
Day 2 of #VapiCon is a wrap!
Two days of incredible conversations with developers, founders, and even competitors.
Big takeaways:
Multilingual AI goes beyond language support. Accents, dialects, and code-switching matter.
Real-world audio is still messy. Overlapping speakers, background noise, and turn-taking remain tough challenges.
Speech-to-speech vs. cascaded architectures? The debate is heating up!
And on-device AI is getting serious attention, especially around privacy and control.
We also wrapped up with TWO winners at the @Speechmatics booth!
One took home an RTX 5070 from our singing challenge, while a young champion smashed our voice-imitation leaderboard and won a LEGO Technic set!
My biggest takeaway? Voice AI isn't just about building better models. It's about the developers, the community, and what we build together.
Until next time, VapiCon!
#VoiceAI #Speechmatics
Speechmatics retweeted
Your voice agent has a listening problem. Speechmatics is working on it.
Our cover story on @Speechmatics
yespress.io/speechmatics?utm… via Yespress
Speechmatics retweeted
Drop by @Speechmatics booth at @Vapi_AI Con 26.
Thank You @Speechmatics for motivating me to speak on camera.
#Speechmatics
Replying to @Speechmatics
@Speechmatics
Come speak to Edgar's services and find out about the best speech text services. He's super friendly, knowledgeable, and can't wait to talk to you.
Speechmatics retweeted
Was just at @Speechmatics booth in Vapi con. Singing to their STT model. It works really well!!!
11 languages tested. 11 first-place finishes. 🏆
Oak captures medical terms across 55+ languages.
See it in action at #HLTHUSA. 🎙️
Live demos + goodies at Booth #V-4025.
Book a meeting here: calendly.com/speechmatics-ev…

ALT The image features a promotional graphic for Speechmatics' new medical model, Oak-1. The design showcases interconnected flags of several countries: France, United States, Lebanon, and Algeria. A tagline reads, "One conversation. Multiple languages. For clinical conversations that switch languages mid-sentence." The Speechmatics logo is at the bottom. The background is a gradient of blue hues.
162
Speechmatics retweeted
Day 1 of #VapiCon is a wrap!
Spent the day talking with developers and founders building everything from restaurant ordering agents to healthcare, recruitment and construction AI.
A few things stood out:
Accuracy in real-world conditions is still a huge challenge. Turn-taking and overlapping speech remain difficult. And more developers want control over their voice AI stack.
We are also seeing conversations shift towards real benchmarks: P50/P95 latency, multilingual accuracy, diarization and testing on actual customer audio.
My biggest takeaway? The next leap in Voice AI isn't just about making agents talk better. It's about making them listen better.
Great conversations, brilliant builders and plenty to explore on Day 2!
#VoiceAI #Speechmatics
Speechmatics retweeted
The booths ready. Looking forward to seeing what everyones building.
I am at #VapiCon at Fort Mason with @Speechmatics today. Bring your demo, come talk to me, tell me what you are stuck on, and come say hello. And win some prizes!!
The best conversations about voice AI tend to happen when everyone gets around the same table. 🍽️🎙️
Great to catch up with the @roarkai_ team in San Francisco ahead of #VapiCon, swapping perspectives on where voice agents are heading and the challenges still worth solving.
The conversations continue at Booth 14 today.
1
1
296
Booth 14 is taking shape. 🔧👂
Final preparations underway for #VapiCon.
Tomorrow, we're putting voice agents' listening skills to the test with live demos.
Come see what your agent really hears.
vapicon.ai
1
209

























