Deepgram
VerifiedDeepgram is a voice AI platform that provides APIs for speech-to-text, text-to-speech, and voice agents. It's built for developers and enterprises that need real-time, accurate transcription and natural-sounding speech generation at scale, covering use cases like transcription, conversational AI, and speech analytics across industries such as contact centers, medical, pharma, restaurants, and government. The platform supports 60+ languages, offers models tuned for different needs (Flux STT for real-time conversation, Nova-3 for production transcription), and includes audio intelligence add-ons like summarization and sentiment analysis. It can be deployed via Deepgram's cloud, regional endpoints, or self-hosted.
Deepgram at a glance
- What it is
- Deepgram is a voice AI platform that provides APIs for speech-to-text, text-to-speech, and voice agents.
- Who it's for
- Aimed at makers and small teams exploring the Speech Recognition & Transcription category.
- Primary use case
- Power enterprise voice solutions with Deepgram’s Speech-to-Text, Text-to-Speech, and Voice Agent APIs.
- CategorySpeech Recognition & Transcription
- PricingFree Tier
- Starting priceFree, $200 credit then pay-as-you-go
- Credit cardNo credit card required
- Foundation modelNova-2
- Bring your own keySupported
- RatingNo reviews yet. Be the first to rate.
Video Spot
Overview
Deepgram is a voice AI platform that provides APIs for speech-to-text, text-to-speech, and voice agents. It's built for developers and enterprises that need real-time, accurate transcription and natural-sounding speech generation at scale, covering use cases like transcription, conversational AI, and speech analytics across industries such as contact centers, medical, pharma, restaurants, and government. The platform supports 60+ languages, offers models tuned for different needs (Flux STT for real-time conversation, Nova-3 for production transcription), and includes audio intelligence add-ons like summarization and sentiment analysis. It can be deployed via Deepgram's cloud, regional endpoints, or self-hosted.
Power enterprise voice solutions with Deepgram’s Speech-to-Text, Text-to-Speech, and Voice Agent APIs. Real-time, accurate, and built for scale.
In detail
Deepgram is a voice AI company offering a unified platform of APIs for speech-to-text (STT), text-to-speech (TTS), voice agents, and audio intelligence. It is aimed at developers and enterprises building products that need to listen to, understand, and respond to spoken language in real time, such as contact centers, transcription tools, and conversational AI agents.
The platform works by giving developers API access to Deepgram's speech models. Flux STT is designed for real-time conversational speech recognition with turn detection and low latency, while Nova-3 is built for production-grade transcription across 60+ languages with strong noise handling. Flux TTS generates conversation-native speech for voice agents, and the Voice Agent API combines STT, a managed language model, and TTS (or lets customers bring their own LLM or TTS) under one billed-per-minute rate. Audio intelligence add-ons like summarization, topic detection, sentiment analysis, and intent recognition can be layered on top of transcription.
Deepgram is used by developers building conversational AI, speech analytics, and transcription products, and by industries including contact centers, medical, pharma, restaurants, and government, where specialized vocabulary and compliance matter. The company states it supports SOC 2, HIPAA, GDPR, CCPA, and PCI compliance, along with a dedicated EU endpoint for data residency.
What differentiates Deepgram, according to its own claims, is having STT, TTS, and voice agent capabilities under one backend, reducing integration gaps and latency between separate tools. It states industry-leading accuracy and speed figures, including faster inference and lower word error rates compared to competitors, along with flexible deployment options (cloud, regional endpoint, or self-hosted).
Pricing is usage-based across all products, billed per second of audio, per 1,000 characters, or per minute of connection time. A free tier gives $200 in credit with no credit card required, a Growth tier offers pre-paid annual credits starting at $4,000/year for a discount, and an Enterprise tier offers custom pricing for large-scale or specific deployment needs.
Common uses
- Building real-time voice agents that can listen, respond, and handle turn-taking in conversation
- Transcribing pre-recorded or streaming audio for contact centers and customer service analytics
- Adding spoken voice output to chatbots or LLM-based applications
- Powering speech analytics such as sentiment and intent detection on call recordings
- Supporting specialized transcription needs in medical, pharma, legal, or finance industries
- Building accessibility or captioning features into applications
Pros
- Combines speech-to-text, text-to-speech, voice agent, and audio intelligence under one API platform and backend
- Free tier with $200 in credit and no credit card required for prototyping
- Supports 60+ languages with models tuned for specific industries
- Flexible deployment options including self-hosting and regional/EU endpoints for compliance
- States enterprise compliance certifications including SOC 2, HIPAA, GDPR, and PCI
Cons
- Pricing is usage-based and can become complex across multiple products, models, and add-ons
- Enterprise-level needs require contacting sales rather than transparent self-serve pricing
- Primarily aimed at developers integrating via API, not a no-code tool for non-technical users
- Growth plan requires a $4,000+ annual pre-paid commitment to get discounted rates
Generated from the listing's details and not yet checked by a Zoipi editor.
Key Features
- Speech-to-text API: Flux STT for real-time conversational transcription and Nova-3 for production-grade transcription in 60+ languages
- Text-to-speech API: Flux TTS, Aura-2, and Aura-1 models for generating natural, conversation-native speech
- Voice Agent API: combines speech-to-text, a managed language model, and text-to-speech, billed per minute of connection time
- Audio intelligence add-ons: summarization, topic detection, sentiment analysis, and intent recognition billed per token
- Custom vocabulary: keyterm prompting and tuned models for industry-specific terminology
- Flexible deployment: run in Deepgram's cloud, through regional endpoints, or self-hosted for data residency and compliance needs
- Language support: 60+ languages with multilingual models and mid-utterance code-switching detection
Best For
- Specialists who want a dedicated speech recognition & transcription tool rather than a general assistant
- Anyone testing the space without an upfront budget
- Buyers who want a highly rated, well-established option
Pricing Model
Free Tier. Generous ongoing free tier you can start with, no credit card needed. Paid plans unlock higher limits.
Pricing Breakdown
Screenshots
Live preview of Deepgram's website.
Frequently asked questions
Community Reviews & Discussion
0 reviews · replies are moderated before they appear.
- No reviews yet — be the first to review Deepgram.