Deepgram
@DeepgramAI
Followers
9K
Following
2K
Media
630
Statuses
2K
Build with one flexible voice AI API. Get started with $200 in credit for transcription or text-to-speech: https://t.co/ssASTgz2u1
San Francisco, CA
Joined February 2016
2024 was a BIG year for us ending with:. 🔸400+ enterprise customers.🔸3.3x YoY growth in product usage.🔸Positive cash flow!. What’s driving our success? (see 🧵). We're building on this momentum as we accelerate into the era of enterprise Voice AI! 🚀.
2
3
6
Introducing Nova-2, our next-gen model for superhuman speech-to-text. TL;DR Nova-2 delivers:.💥 Next-level accuracy: +18% accuracy than Nova-1 & over 36% accuracy than OpenAI Whisper large.💥 Up to 40x faster.💥 Same low cost: 3-7x cheaper. 🧵👇
72
223
1K
The wait is finally over! We're thrilled to present Aura text-to-speech to the world. 👋. TL;DR it checks all the boxes:.⚡️ Lightning-fast: <250 ms latency.🗣️ High-quality voices: natural tone, rhythm, emotion.🚀 Cost-effective for high-throughput use cases
46
100
506
We’re thrilled to introduce Aura, a powerful text-to-speech (#TTS) model built for real-time, conversational voice AI agents. 🎉. Check it out: Let's dive in. 🧵.
51
102
442
Introducing our unified Voice Agent API! ⚡️. Quickly and easily build voice agents that listen, think, and respond in real-time, as naturally as a human can. Powered by the industry's fastest, most performant speech recognition and voice synthesis models to deliver:
33
89
414
The wait is over! With 60M+ minutes transcribed, our next-gen speech-to-text model Nova-2 is now available. What's new?.✅ Expanded languages: Spanish, Hindi, German, French, Portuguese.✅ Custom model training .✅ On-prem deployment . Let's dive in. 🧵
39
85
345
Learn how to build a voice AI agent from the ground up in our virtual workshop with @GroqInc. -Hands-on coding and expert guidance .-Best practices for responsive voice interactions .-$1K in Deepgram credits to build with. 📆 Sept 20 @ 9AM-12PM. Register:
28
79
277
We’re thrilled to announce a major expansion of Deepgram’s Audio Intelligence lineup. 🚀. Introducing new models for sentiment analysis, intent recognition, and topic detection for enhanced understanding of conversational interactions. Let's dive in. 🧵
14
29
160
MAJOR NEWS🚨 We're excited to introduce Nova⚡️ our most powerful speech-to-text model. TL;DR it surpasses competitors on:.💥Speed: 23x faster.💥Accuracy: 22% fewer errors .💥Cost: 3-7x cheaper. Let’s dig in on how Nova outperforms every ASR model… 🧵
18
35
183
Healthcare's $1 trillion admin problem has a solution: AI. Companies like @chartnote , @Sonia_Health, and @StackAI_HQ are revolutionizing care, automating note taking, EHR documentation, QA, and more. We're powering health tech apps with our Nova-2 model for real-time medical
7
21
134
Work at the speed of voice. 🚀. Introducing Shortcut by Poised, a revolutionary Voice AI tool that finally brings the natural power of human conversation to your daily work!. 🎉 We're live on Product Hunt today! Support us here:
10
14
75
In this week's Twitter Space, @BekahHW will chat with @TessaMero about doing DevRel during Hacktoberfest. You can set your reminder now! And if you have questions you want to hear answered, you can drop them in our discussion:
4
11
50
We completed our $72M Series B with $47M in new funding to define the future of AI Speech Understanding 💥. Details here: Or keep reading 🧵.
2
10
41
Huge shoutout to Jonathan Ross at @GroqInc for this incredible LUI demo on @CNN! We're honored to be a part of your stack & we can't wait to see where this astounding technology will go next 🚀. Thanks again @JonathanRoss321 for the mention 😄. Demo:
2
6
36
📢 Twitter space alert 📢.Who: @BekahHW and @lindseybieda .What: @github discussions.When: Friday, Sept. 23, 12p PT.
1
5
26
Speech-recognizing AI, Raspberry Pi, and a wearable screen mounted chest-high. All you have to do is wear a hoodie that records your voice—and let Deepgram's #speechtotext take care of the rest. @zackfreedman. #captioning #voiceAI.
0
3
24
We've got another Twitter Space coming at you this Wednesday at 1pm ET. @BekahHW and @saltnburnem will be talking about Getting Started with Accessibility. Set your reminder here: Ask questions in advance here:
0
3
24
We're so pumped about Hacktoberfest, we have Twitter Spaces coming at you every Wednesday for the next three weeks! We start tomorrow at 8a PT with @nickytonline, @KirkCodes, and @BekahHW chatting about Getting Started and answering your questions.
0
6
18
NEW: Next-gen speaker diarization and language detection models trained on 100K+ voices!. TL;DR we now offer:.🟢 Language-agnostic diarization that’s 10X faster vs. next competitor.🟢 Revamped auto language detection with <4% error rate. Let's dig in 🧵.
2
5
16
Hello, hello! This week @BekahHW is taking over the Deepgram socials🎉 . I'm the Technical Community Builder on the DevRel team, which means that I help to educate developers on how to use our Speech Recognition API, while focusing on community interactions.
1
2
15
After playing with Whisper last week, we made it easier to try via the Deepgram API. As easy as copying and pasting, it takes just 1 minute to get started (you're welcome). 😎 . #OpenAI #OpenAIWhisper.
1
2
14
Developers now can build ultra-low latency voice-to-voice AI with Daily Bots! We’re excited to partner with @trydaily to be the default speech-to-text engine for Daily Bots, the open-source cloud for ultra-low latency voice and video agents.
Today we’re launching Daily Bots, the ultra low latency Open Source cloud for voice, vision, and video AI. Build voice-to-voice with any LLM, at conversational latencies as low as 500ms. With Daily Bots, developers can:. *️⃣ build with Open Source SDKs.*️⃣ mix and match the
1
1
15
As a bootcamp grad, I'm a big fan of projects that help early career folks to grow and find their way into their first roles. We've got a couple of blog posts to help folks upgrade portfolio projects as they work their way to their first tech job 🧵.
1
3
13
That's a wrap for Deepgram at #kcdc2022. Thanks to everyone who came by the booth or heard @the_BrianB and @BekahHW give their talks. See you next time!. Sidenote: Deepgram loves @rustlang but our PM Shir is taking it to a new level with that hat.
1
0
13
We're back for day 2 of Project Voice! 🤩 We're excited to meet everyone and hear about all the latest in #voicetechnology. 🚀. #voiceAI #voicetech #voicefirst #speechrecognition
1
2
11
ICYMI, @zackfreedman designed a hoodie that transcribes his every word using Deepgram #API. Tune in to learn how he did it and see people's live reactions in public:.
1
3
11
Hello Twitter - @_phzn here! I’ve somehow wrangled control of this account for the week - thought I’d take a min to say hi 👋🏻 I’m a Developer Advocate here, which means I helps devs discover and learn how to use our Speech Recognition API!
1
0
13
Word is that @BekahHW killed it at @kc_dc talking about "The Traps of Learning Tensorflow.js and Why You Should Learn It Anyway". Catch her at the Deepgram booth or live podcasting with the @VirtualCoffeeIO gang!. #kcdc2022 #TensorFlow #machinelearning.
1
1
12
It's not too late to get involved in our Hackathon with @ThePracticalDev! . 🛠 Build an app using Deepgram.📝 Tell us about an interesting use-case for Deepgram.📣 Cheer participants on . And win some cool prizes 🥳 Head to the link below to learn more!.
0
1
11
Hey, hey! Say HELLO to Julia Strout, @sandra_rodgers_ & @tonyasims from @DeepgramAI at #SciPy2022 …and get a free polaroid at our photo booth. 📸
1
3
11
Join @BekahHW and @codergirl1991 tomorrow as they talk about being a maintainer during Hacktoberfest, share tips for contributors, and take all your questions!.
0
4
11
We highlighted how amazing LiveKit is before, but their work is so incredible it bears repeating: . It’s now possible for anyone to build real-time server-side agents that leverage our AI alongside many other models. A fantastic feat of engineering & we’re proud to be part of it.
When AI is as smart as a human, we’ll interact with it like we do with each other. Human interaction is real-time and multimodal:.We use → 👀👂👄.AI uses → 📹 🎙️🔈. Today we're launching a stack for building real-time multimodal AI apps.
0
5
12
Who: @BekahHW & @JuliaSeidman.What: Twitter Space, Developer Growth: Writing to Learn. When: Nov. 9th, 10a PT.Where: Have questions you want to be answered? Drop them in our discussion:
1
1
11
Want high quality, open-source speech-to-text AI? Check out our new captioning packages!. For JavaScript and Python, these OSS packages empower developers to effortlessly generate SRT and WebVTT captions!. Learn more below 👇
1
2
10
We're in good company on the #AI100 list. Thanks @CBInsights for the recognition for our audio search technology! #CBISummit
0
2
10
Our Head of Developer Relations @baldbeardbuild will be live on #SignalTV tomorrow to show viewers how to add live #captioning to a #TwilioVideo room using AI & deeplearning. If you plan on attending Signal, be sure to tune in! 📺. @deepgramdevs #SIGNALCONF #Twilio.
0
1
10
We’re excited to announce that Deepgram has raised a $25 Million Series B from Tiger Global and others to put the power of voice data behind the world’s big ideas. Read all about it here 👉
0
1
11
Deepgram has transcribed over a trillion words since we’ve launched! That’s about the same as. - 22,000 x the words in the entire Encyclopedia Britannica.- 1,100,000x the size of all Shakespeare’s works.- 1162x the number of words you’ll speak in your lifetime!
1
0
9
Wow, those response times were really fast! We’re so excited that our API helped you minimize latency so effectively. Thanks for the amazing demo! 🔊🚀.
Shayne and I built an insanely fast AI voice assistant in 50 LOC. Llama 3.1 running on @CerebrasSystems. 2.5x faster inference than literally anything else. 🔥 400ms response times. Uses:.🌐 @livekit transport.👂 @DeepgramAI STT.🧠 @CerebrasSystems LLM.🗣️ @cartesia_ai TTS
0
1
11
Hey @elonmusk. A project at our internal hackathon is a game built with @reactjs, @DeepgramAI, and @tailwindcss. Say quotes, lyrics, or tongue twisters as fast as possible. Can you reach ludicrous speed? It's hardcore and doesn't cost a penny to play.
0
2
11
Nova is not only 22% more accurate and 23x faster, but also the lowest-cost alternative starting at just $0.0043/minute, making it 🤑 3-7x more affordable than any other provider out there.
2
3
10
People ask: How does it work? How does Deepgram transcribe audio? so quick, so speed, so good?. Here's a behind the scenes:
0
2
10
We launched #free #deeplearning powered #transcription for the masses today! check it out! @TechCrunch @JohnMannes.
0
10
9
Shoutout to everyone who's contributing to our Open Source projects during Hacktoberfest!!!. @davidalpert .@rainalldaylong .@nickytonline .John Bowser.@dotaadarsh .Nemrud Demir.
0
0
8
Tune in to our session on Deep Learning Speech Recognition for Conversational AI next Thursday, October 8th @ 12PM PDT (7PM UTC) @ NVIDIA #GTC20 online.
0
4
10
What a beautiful lineup! Very reminiscent of the Avengers 😊.
8 AI tools you dont want to miss 👇👇. 📷 Images: @midjourney.🎬 Video: @runwayml.✍ Writing: @rytr_me.💬 Speech: @DeepgramAI.🎨 Design: @diagram.🚀 Marketing: @copysmith_ai.👛 Sales: @lavenderhq.👍 Support: @CohereAI.
1
1
10
We launched the Deepgram Startup Program to put the power of #voicerecognition in the hands of developers. Plus, we released new features: Console, SDKs, and API documentation- all designed to create a frictionless #developerexperience. Read all about it!. #startups #devx #dx.
0
2
10
Do you like our stickers? We do too! Come find @DeepgramAI, a speech-to-text company transcribing real-time and pre-recorded audio at #SciPy2022 this week to snag some!
2
0
9
We're firm believers that voice will transform the way humans interact with tech. And this shift starts with human-like AI. That's why we created Aura - our incredibly fast, natural-sounding, and cost-effective text-to-speech API. ✨.
2
2
10
There are now have AI resources for 4 out of 5 senses: Sight, hearing, taste, and smell. Wanna see them in action? Check out the links below! 😊.
3
2
10
Now available: Aura WebSocket Text-to-speech API to support fast input streaming, ensuring your AI agents are more responsive than ever! 🚀. TL;DR:.-3x faster than ElevenLabs Turbo 2.5.-70% time savings in LLM to TTS latency with token-by-token transmission.-Supports 40+
2
0
10
admiring the fantastic @_phzn for the work they do for @DeepgramAI (including the #wearable they built for @emfcamp).an example of #DevRel at it’s finest!.as well as the incredible things they do with @yougotthisconf.listen to the full episode here:
0
2
9
Thanks to everyone who met with us at @NVIDIAGTC! We talked to sooo many of you about the future of AI and the world at large!. As a closing treat, check out this video of our very own CEO @deepgramscott playing Yappy Bird! (aka our vocally-controlled Flappy Bird demo) 🐥
0
1
7
Introducing our new and improved Nova-2 Medical model that delivers:.🎯Superior accuracy: 43% WER improvement vs. benchmarked alternatives.⚕️Enhanced recognition of medical terms: 20% higher accuracy vs. leading alternatives. See the benchmarks. 🧵
6
0
9
📣 We're excited to announce our partnership with @trydaily on the beta launch of their newest #API, which lets developers add live Deepgram #transcription to a Daily call with upwards of 90% accuracy. Learn more in their latest blog:.
0
1
8
Is #transcription just part of a broader project? Learn how to #translate real-time speech or pre-recorded audio from our latest blog tutorial from @_phzn. #developertools #STT #voiceAI.
0
1
7
Extending upon Nova's groundbreaking training, which spanned +100 domains and 47 billion tokens, Nova-2 continues to be the deepest-trained ASR model in the world.
1
1
9
Our friends at @trydaily featured this incredible blog+demo about talking to an LLM with your voice! . Shoutout to them for using our latest speech-to-text AI in their fantastic product 🚀. Here's the link:
2
2
9
The World’s Fastest Voice bot was created using @trydaily, @cerebriumai, @DeepgramAI, and LLaMA. Here’s how it works:. Link:
1
3
9
When it comes to Machine Learning and coding languages, Python seems to be the head honcho. However, other languages have AI libraries and packages as well! Here’s a list:.
1
0
8
📈Consumer vs. Enterprise:. OpenAI’s API: We think OpenAI’s approach is more consumer-focused, (and, hot-take) likely built by chaining together models like whisper-large-v3-turbo, gpt-4o, and their 6 TTS voices for smooth, natural interactions. (Why? High cost + use of a VAD).
1
0
9
We are expanding early access to a select group of builders to integrate with Aura as we continue to fine-tune and refine the model and production API. Learn more about Deepgram Aura, or join the waitlist to be the first to try it out!. Full announcement:
2
0
8
This video tutorial is amazing! Thanks so much for creating and showcasing your project. We’re glad we could be a part of your progress 🥳🚀. P.S. We’re a big fan of your Nova notebooks! We’ve got a couple of our own in our docs, if you want to check them out 📓.
How to use OpenAI's Whisper and get word-level timestamps using @DeepgramAI 's API?. And how to generate a subtitle file (.srt) file with word-level timestamps and create Instagram-style fast word-scrolling text videos? . Video tutorial coming soon! . But the code and easy-to-try
0
1
9
< 30 second coding tutorial!. Download audio from *any* YouTube video you want with a few brief lines of Python 🐍. Check it out!
1
1
7
What's more valuable: #algorithm or #dataset? @deepgramscott and @susansoptimum talk it out in the AI Show #podcast:
0
2
9
It’s an honor and a pleasure to provide the 🔥🔥🔥 to SWCs like you!.
been using @DeepgramAI for real time voice streaming on next js and its so fire (and easy to use).
1
0
8
We raffled off a Nintendo Switch at #SciPy2022. Congrats to our grand prize winner, @ATHARVBHAT! Happy gaming! 👾
1
1
8
📣 Employee spotlight alert! 📣 This week, we're celebrating one of our newest Deepgrammers @TonyaSims! . We're excited to have you aboard, Tonya! 🎉 . Check out our careers page for all open roles:.
1
0
6
Y Combinator has funded more than 4,000 startups! We are honored and proud to be among YC’s top companies for 2023! Check out the full list here:
2
1
8
Deepgram's model discovery lead, Andrew Seagraves, examined #opensource speech recognition models and their performance characteristics.
2
0
7
Exciting day for us here at Deepgram! Today, we announced $12 million in Series A #funding and two new platform features. Read more about our news and how we built our #speechrecognition tool from the ground up on @techcrunch via @alex:
0
0
8
Our employee spotlight this week is Luca Todd! 😎 . Do you have what it takes to join our team of rockstars? Apply now: #DGisHiring #voicetechnology #voiceAI
2
0
7
"Our mission at Deepgram is to power the next generation of voice experiences, and we know prioritizing our people is critical to achieving that vision" -@deepgramscott . ➡️ Learn more: ➡️ Join our team:
1
1
8
Remember when @_phzn built a wearable live captioning badge? See how you can build yours in his latest YouTube tutorial 😎🔥. #transcription #speechrecognition #STT #voiceAI .
0
2
7
With our latest speech understanding feature, #Summarization, you can automatically generate relevant summaries for your audio data. More below 👇.
0
0
7
Live from #ProjectVoiceX, our CEO @deepgramscott talking about how Deepgram is empowering #developers to take full advantage of #voice in their applications. 🗣.
1
0
7
Liquid neural networks improve upon traditional neural nets in unprecedented ways. With dynamic neurons and a suite of optimizations, liquid neural networks are pivotal to understand to keep up with the AI world. Here’s all you need to know:
2
2
8
No better way to start my takeover week except to say I tried a thing - recreating some of the most famous memes ever with #dalleMini. Attempts in thread. Current theory: AI understands one thing, and it's Shiba Inus. First up - Disaster Girl.
3
1
8
Pipedream + Python = a readable transcript in your inbox for each new episode of your favorite podcasts, thanks to @_phzn .
0
3
7
We love to see it! 👏.
Do you know how accurate @DeepgramAI is at transcribing speech?. Very!. Tonight I learned how to run Deepgram on a downloaded podcast episode. It transcribed all of @theChrisDo and Daniel Priestley talking about branding. Now to dive in and take notes 📝
0
0
6
Well, friends, it's been a great week. Thanks for being there for my social media takeover of the DG account. Next week the amazing @yujian_tang will be taking over, so stay tuned for some awesome content.
0
1
7
Thank you @ToolboxforB2B for mentioning us as a #Top10 Speech Recognition Software for 2022! 🤩. #voiceAI #voicefirst #speechtotext #transcription .
1
3
7
When it comes to understanding transcripts, punctuation can be a real lifesaver. Here's why punctuation and capitalization matter for speech-to-text. #speechtotext #transcription
0
1
7
Read the detailed breakdown of Nova's performance and benchmarking here:
0
2
8
I've been using @pipedream for a couple months now and it is seriously sweet. @_phzn wrote up a how-to on transcribing files in your @googledrive automatically with PipeDream and Deepgram. #lowcode #javascript #speechtotext.
0
1
6
A Python Machine Learning Engineer 🐍💻 is sent on a special mission: Operation Topic Detection. Your task? Transcribe a podcast with topics ASAP. Know how to solve it? Our latest post from @tonyasims. #podcasttranscription #PodcastMovement #Python.
1
0
6
Wanna see how to use automated speech recognition for fun? Check out these videos! 🎥📺. Subtitles in Real Life: . Voice controlled car: Coding a website with voice-only:
0
1
7
Nova's groundbreaking training spans over 100 domains and an astounding 47 billion tokens, making it the deepest-trained automatic speech recognition (ASR) model to date.
1
0
7
What’s the current state of #ConversationalAI? Join our virtual panel with experts from @Bitext and @OneReach to find out! Plus, learn about the necessary components to building exceptional customer service experiences. Register below! . #AI #voicebots.
0
2
7
We are honored to be selected as a 2022 Silver Stevie Winner in the "Sales & Customer Service Solutions Technology Partner of the Year" category for @TheStevieAwards! 🏆. Thank you to our incredible team and all the work you do for our customers! .
0
1
7
We're stoked to be here for day 2 of the Conversational Interaction Conference! Come say hi to our team 👋 — and don't forget to grab a hot sauce sampler of our Smokin' Hot ASR! 🌶. #CICON22 #voiceAI #speechrecognition
2
0
7
0
1
5