The Browser Agent SDK is here. 🚀 Four composable npm packages that drop a Deepgram voice agent into any web app. → @deepgram/agents-widget: drop-in widget, six layouts, no framework required → @deepgram/ui: pre-built React components with CSS-variable theming → @deepgram/react: provider and hooks for state, audio, and function calling → @deepgram/agents: framework-agnostic core for Vue, Svelte, Angular, or vanilla JS Same reconnection logic, playback-aware mode tracking, audio buffering, optional Silero VAD, and KeepAlive across every layer. Start with the widget today, graduate to the React layer when you need custom UI, drop to the core when you outgrow React. You don't change vendors as you grow. Ship in minutes, customize for months. 💫
Deepgram
Software Development
San Francisco, California 38,669 followers
Powering the Voice AI economy—trusted by 200,000+ developers to build low-latency, enterprise-grade Voice AI at scale.
About us
Deepgram is the real-time API platform powering the trillion-dollar Voice AI economy. Backed by a $130M Series C at a $1.3B valuation, Deepgram is trusted by 200,000+ developers and 1,300+ organizations to build Voice AI products, platforms, and autonomous agents with the lowest latency, highest accuracy, and enterprise reliability. Our voice-native foundation models and runtime infrastructure have processed 50,000+ years of audio and over 1 trillion words, making Deepgram the most experienced voice AI platform in the world. Industry-leading models & platform: 👂 Nova-3 — the world’s most accurate real-time speech-to-text model 🔊 Aura-2 — professional, enterprise-grade text-to-speech 💬 Flux — the first Conversational Speech Recognition model designed to handle interruptions 🚀 Voice Agent API — enterprise-ready, real-time conversational AI 🧠 Saga — the Voice OS Beyond core infrastructure, Deepgram is expanding the Voice AI ecosystem through: 💪 Powered by Deepgram, supporting voice products built by leading AI startups and enterprise organizations 🌉 A new Voice AI Collaboration Hub in San Francisco for builders, partners, and the voice community 🍔 The acquisition of OfOne, delivering real-time Voice AI for restaurants and drive-thru operations with 95%+ containment 📃 A growing patent portfolio in Voice AI Much like APIs powered the payments and cloud economies, Deepgram is building the foundation for a trillion-dollar B2B Voice AI economy—centered on the most natural human interface: voice.
- Website
-
https://deepgram.com
External link for Deepgram
- Industry
- Software Development
- Company size
- 51-200 employees
- Headquarters
- San Francisco, California
- Type
- Privately Held
- Founded
- 2015
- Specialties
- Speech Search, Transcription, Speech Recognition, Audio Understanding, Speech Analytics, Voice Recognition, Artificial Intelligence, Deep Learning, Natural Language Processing, Text-to-speech, Voice Generation, and Conversational AI
Products
![Click here to view Deepgram]()
Deepgram
Speech Recognition Software
We created a powerful transcription and speech understanding API built for developers. Our AI technology uses real-world data for consistent improvement, providing unmatched speech-to-text accuracy that easily outperforms our competitors. Our end-to-end deep learning offers unrivaled ASR solutions with superior speed, real-time transcriptions, and scalability. That's why leading startups to enterprises trust Deepgram.
Locations
-
Primary
Get directions
548 Market St.
Suite 25104
San Francisco, California 94104, US
-
Get directions
207 E Washington St
Ann Arbor, Michigan 48104, US
Employees at Deepgram
Updates
-
A Fortune 50 retail pharmacy scaled its Interactive Voice Response (IVR) to 1M+ calls per day across 7,000+ locations with Deepgram. 🏥 📈 92% recognition accuracy on pharmacy vocabulary 🧠 85.3% on complex medical terms 📞 14,000 calls/hour in production ✅ Higher containment, fewer misroutes, lower agent costs. They replaced a legacy Nuance IVR with Nova Medical (STT) and Aura (TTS). The architecture supports multilingual expansion and new workflows without replatforming. Full story: https://lnkd.in/d8YJT2Hq
-
-
Nova-3 expands speech-to-text support across Asia-Pacific. 🌏 New support now available for: 🔹 Thai 🔹 Cantonese Traditional 🔹 Mandarin Simplified 🔹 Mandarin Traditional We’ve also improved speech recognition accuracy for Bengali, Marathi, Tamil, and Telugu, and added new support for Gujarati. Built for production-ready voice AI across streaming and batch workflows. https://lnkd.in/gZ4nWb-W
-
Today we're announcing Flux Multilingual for Restaurants. 🌎 🍽️ Built for restaurant Voice AI, it handles natural code-switching across 10 languages without a perceptible latency hit. In a diversifying world, brands that force customers to order in a language that isn’t theirs risk lower customer satisfaction, weaker brand trust, incorrect orders, and slower service. Restaurants can now deliver faster, more accurate multilingual interactions across ordering, support, and service experiences, in the language each customer is most comfortable in. Available in English, Spanish, French, German, Hindi, Russian, Portuguese, Japanese, Italian, and Dutch. One model. One API. Monolingual-grade accuracy. Learn more: https://lnkd.in/gMK-zkgH
-
-
Big congrats to Vapi on their $50M Series B at a $500M valuation! Vapi makes it easy to build voice agents with people skills. ✅ 2.5M+ agents launched and counting ✅ 750K+ developers from startups to Fortune 500 ✅ 1 Billion calls All powered by Deepgram's leading voice APIs. "Our voice capabilities powered by Deepgram APIs deliver the accuracy, low latency, and quality our customers need to build reliable, real-time voice agents." — Nikhil Gupta, Founder and CTO at Vapi #PoweredbyDeepgram
-
Deepgram reposted this
🤔 Have you tried Deepgram Velocity yet? We've just published a new version, with a couple of UX enhancements! Run Velocity at Windows login, and choose whether to insert text into the currently focused app, or the app that had focus when you started recording. Velocity is an #opensource dictation utility for Windows 11 users, enabling you to speak to your AI agents and write documents with just your voice! 🎙️ Grab the pre-built release here, or compile it yourself ➡️ https://lnkd.in/gjVVs4Ci Check out the video overview 📺 https://lnkd.in/gyx5zxYp Sign up for a Deepgram account, and get $200 in credits to use for STT and TTS! 💳 No cc required to use credits, but will be needed to add more credits. #Deepgram #VoiceAI #software #Rust #Rustlang #Transcription #Dictation
Velocity 0.5.0: Launch at Startup & Set App Focus 🎙️ Dictation for Windows 11 Users
https://www.youtube.com/
-
Last week, we launched Flux Multilingual, one conversational speech model for global voice agents. Now let’s look at the benchmarks 👇 We benchmarked Flux Multilingual on real-world production audio across all supported languages using each vendor’s default streaming configuration. 🏆 Result: Best-in-class WER across the majority of supported languages, including: 🇺🇸 English 🇪🇸 Spanish 🇩🇪 German 🇫🇷 French 🇵🇹 Portuguese 🇮🇳 Hindi Conversational latency matters just as much as transcription accuracy. Across aggregate end-of-turn benchmarks, Flux Multilingual delivers: ⚡ Highest aggregate EoT F1 across all supported languages ⚡ Up to 3x lower latency than competing real-time EoT systems Flux doesn’t rely on silence thresholds, it models conversational context directly. For voice agents, that’s the difference between reacting to silence and understanding conversation.