Speak '26 registration is officially open! 🔥 Voice AI is on the verge of its biggest milestone yet. On Oct 29, the people building it are getting together in SF to talk about what it takes to get there. Register: deepgram.com/speak

Voice AI is growing because builders keep pushing it forward, and we believe in backing teams early. On top of the Deepgram Startup Program, we're proud to announce that we're a launch partner in @livekit's Startup Program. Pre-Series A teams get $1,000 in Deepgram credits and it works natively through LiveKit Inference. Ready from your first line of code and as you scale. Can't wait to see what you build: livekit.com/startups

Some of the biggest Voice AI opportunities live in places where the cloud isn’t always an option. That’s why Deepgram is excited to support @VAST_Data's launch of DataEnclave – bringing leading AI models into customer-controlled infrastructure while helping protect sensitive data and models in use. Deepgram already supports self-hosted Voice AI for enterprises that need greater control over where their workloads run. With VAST DataEnclave, those teams have another path to put real-time Voice AI to work in the environments that matter most. Because the best AI infrastructure meets enterprises where they are. More on DataEnclave: bit.ly/4hhYt4T

Seeing so many cool @typesafeai builds! If you're using Jev and @DeepgramAI in your build together, tag me so I can see!!!!

So look who made it to @HackMIT! 🙌🏾 I’ll be here building and talking about interesting use cases when you add voice to applications. And @DeepgramAI brought some extra motivation: Nintendo Switch 2s and some really good cookies!

Quick update - I’m taking @Amtrak up to @HackMIT this weekend! Bringing some @DeepgramAI credits and cookies with me too 😊😊😊

I spent a few years working on dark matter detectors, including the LUX-ZEPLIN (LZ) experiment. At the time, I definitely didn’t imagine that experience would eventually lead to a company building voice AI. It’s exciting to see the latest result from LZ getting a new signal to chase, but the bigger thing is what work like this creates around it. When you try to answer a really hard scientific question, you have to build new instruments, new simulations, and new ways to work with massive amounts of complex data. That work doesn’t just stay inside the original experiment. It creates capabilities that go in directions nobody expected. That’s what happened with @DeepgramAI. We were working with waveform data in particle physics and trying to understand it. That led us to think differently about how machines could process complex signals at scale. We weren’t trying to build voice AI, but eventually that’s where the work took us. That’s what I find cool about moments like this. The discovery matters, obviously, but so do all the things that become possible because people decided to go after a hard problem in the first place. Congratulations to the LZ team, excited to see what comes next…

I will be down in San Jose next week for the @WeAreDevs World Congress! Come find me at @DeepgramAI's booth #752 and say hello? 🩷🩷🩷

first real miss of the day: "we should go left on the pricing page" came back as a maze move. left, one square, 0.99 confidence directional language is everywhere in office speech in a way that "jump" and "duck" (probably) are not

PIVOT! ok so single words are not great for end of turn detection, the games working, but not the best experience. lets go with a maze navigator instead

Proof of concept: Robo Jumper To keep our little robot alive, we gotta make sure we're giving correct instructions to avoid obstacles. The bad news for this little fella: I'm in the same room my wife is taking a meeting... So much corporate speak to ignore.

Stand back everyone. I figured out how to use my @elgato stream deck to do camera moves. @DeepgramAI "What is Flux-TTS" video series incoming!

Jev is a strong fit for voice driven interaction layers with @DeepgramAI in busy spaces. Flux "eager" turn detection fires events before you are done talking, and jev's fixed option set goes at the question a room makes hard: "was that meant for me, or the person behind me?"

Cool demo by @syntax showing @deepgramai working with @typesafeai Jev: Deepgram accurately transcribes (STT) the raw audio with NER (via keywords), an LLM generates notes and a summary from that transcript, then Jev then fact-checks the result rather than trusting it blindly. Jev extracts topics, tags, and chapter markers, then verifies each individual claim the LLM made to catch hallucinations. Crazy to see how many voice driven applications possibilties there are with this technology. Video: piped.video/QbYBRjOaGOo?t=615 Source: github.com/syntaxfm/website/…

Your whole voice stack, free for a year on LiveKit. $10K in Cloud credits + a year of Scale + $7K in partner inference credits ($1K each: @googlegemma, @DeepgramAI, @inworld , @rimelabs , @FishAudio, @AssemblyAI, @GradiumAI.) One workspace. No provider keys. Apply now: livekit.com/startups