Deepgram
FreemiumReal-time STT, TTS, and Voice Agent APIs for developers.
By Tanmay Verma, Founder · Last verified 06 Jul 2026
6.3k views
Added 4/3/2026
95/100Safe BetIn short
Deepgram — Real-time STT, TTS, and Voice Agent APIs for developers. Best for Developers building real-time voice agents and conversational AI, Contact centers needing live transcription and agent assist, Healthcare providers automating medical transcription at scale. Free to start; paid plans from $4/mo.
Is Deepgram actually worth it?
LiveSee what real users actually say. We scan live discussions, reviews and complaints across the web and hand you an honest verdict — in under a minute.
3 free scans · no card needed · downloadable report
Editorial Verdict
Best for
Developers building real-time voice agents and conversational AIContact centers needing live transcription and agent assistHealthcare providers automating medical transcription at scaleGlobal apps requiring multilingual speech recognition in one APIPlatforms embedding Voice AI into their products via partner program
Not ideal for
Simple transcription-only use cases (Nova-3 standalone is cheaper)Teams needing an out-of-the-box UI (no built-in frontend)Low-budget hobbyists—free tier is limited to $200 creditOn-premise-only deployments without self-hosting expertiseUse cases requiring real-time video analysis (no video support)
Deepgram's unified Voice Agent API and real-time Flux models make it a top pick for developers building conversational AI. If you need simple batch transcription, Nova-3 standalone is cheaper. Best for those who value a single API over stitching components.
Skip Deepgram if Skip Deepgram if you need a turnkey UI or are only doing batch transcription on a tight budget—the API-first design and per-minute costs may not fit.
Compare with: Deepgram vs AssemblyAI, Deepgram vs ElevenLabs, Deepgram vs Whisper Memos
Last verified: July 2026
What's new in Deepgram
Checked 2 days agoAcross the latest 6 updates: 1 feature update and 5 changelog entries.
FeatureChangelog·7 days agoNewest
llms.txt documentation index available at root level
Added /llms.txt documentation index for AI agent discoverability. Append /llms.txt or .md to any page for markdown versions.
ChangelogChangelog·8 days ago
July 1 changelog entry
Changelog entry for July 1, 2026. No further details provided in input.
ChangelogChangelog·9 days ago
June 30 changelog entry
Changelog entry for June 30, 2026. No further details provided in input.
ChangelogChangelog·22 days ago
June 17 changelog entry
Changelog entry for June 17, 2026. No further details provided in input.
ChangelogChangelog·24 days ago
June 15 changelog entry
Changelog entry for June 15, 2026. No further details provided in input.
ChangelogChangelog·28 days ago
June 11 changelog entry
Changelog entry for June 11, 2026. No further details provided in input.
Viability Score
95/100
Safe BetHow likely is Deepgram to still be operational in 12 months? Based on 4 signals — momentum (how recently it shipped), wrapper dependency, revenue model, and web presence.
momentum
100
funding runway
80
website health
90
wrapper dependency
100
Last calculated: July 2026
How we score →Key Features
- Real-time speech-to-text with Flux and Nova-3 models
- Text-to-speech with Aura-2 and Aura-1 voices
- Unified Voice Agent API (STT+TTS+LLM orchestration)
- Flux Multilingual: conversational STT in 10 languages with built-in turn detection and interruption handling
- Batch transcription for pre-recorded audio
- Self-hosted deployment option
- Audio Intelligence API for emotion, sentiment, topic detection
- Custom model training for domain-specific accuracy
- Speaker diarization and sentiment analysis
- WebSocket and REST API integration
- Profanity filtering across models
- Automatic language detection (Nova-3 Multilingual)
- Keyterm Prompting for domain-specific jargon accuracy
- Redaction of PII like social security numbers and credit cards
- Smart Formatting for punctuation, casing, dates, currency
About Deepgram
FreemiumAdvancedAPI availableAPI
Deepgram is a Voice AI platform that provides real-time and batch APIs for speech-to-text (STT), text-to-speech (TTS), and voice agents. Developers and product teams use Deepgram to build conversational AI, contact center analytics, medical transcription, and voice-enabled applications. Key capabilities include Flux Multilingual (10 languages in a single model with built-in turn detection), Nova-3 for high-accuracy transcription, Aura-2 for TTS, and a unified Voice Agent API that combines STT, TTS, and LLM orchestration into one endpoint. You can deploy on cloud or self-hosted. Compared to alternatives like AssemblyAI or Google Cloud Speech-to-Text, Deepgram offers lower latency and a unified API that reduces integration complexity.
Behind the Verdict
Deepgram is for builders who want one API to handle the full voice interaction loop—STT, TTS, and LLM orchestration. The Voice Agent API cuts integration work dramatically compared to assembling separate providers. Flux Multilingual, launched May 2026, is a standout: it handles 10 languages in a single streaming model with automatic turn detection and interruption handling, which is rare in the market. Pricing is usage-based and transparent; the Pay As You Go tier gives $200 free credit. Growth tier saves up to 20% with annual pre-paid credits. Enterprise gets custom SLAs and self-hosting. We'd pick Deepgram for real-time voice agents, contact center analytics, and multilingual transcription at scale. Where it bites: there's no built-in frontend UI, so you must build your own interface. The free credit is limited, and streaming concurrency caps may pinch high-volume use cases without the Growth plan. Custom model training requires a sales conversation. Compared to AssemblyAI, Deepgram's unified API and lower latency give it an edge for interactive voice, but AssemblyAI's docs and SDKs are more beginner-friendly. For batch transcription only, Nova-3 is competitive but you pay per minute. In practice, the Voice Agent API is the real value—skip it if you only need STT.
Researching Deepgram? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Deepgram actually fits — and what changes day-one when you adopt it.
Developer building a voice agent
You want to create a multilingual customer support voice bot with natural turn-taking.
Outcome: Use Deepgram's Voice Agent API with Flux Multilingual. You get STT, TTS, and LLM orchestration in one endpoint. The built-in turn detection handles interruptions naturally. You can be live within hours.
Contact center analyst
You need to transcribe 10,000 hours of recorded calls per year and analyze sentiment.
Outcome: Use Nova-3 batch transcription (monolingual $0.0048/min) plus Audio Intelligence API. Deepgram's batch processing at scale with diarization and sentiment analysis gives you a full transcript and insights.
Platform partner embedding Voice AI
You want to add voice features to your SaaS product for multiple customers.
Outcome: Join the Deepgram Partner Program for API integration. Use the unified Voice Agent API to add STT/TTS without stitching components. Your customers get a seamless voice interface.
Use Cases
- Build real-time voice agents for customer support with natural turn-taking
- Transcribe live meetings with speaker labels using Nova-3
- Analyze call center recordings for sentiment and compliance
- Generate captions for video content with low latency
- Create multilingual voice assistants with Flux conversational STT
Models Under the Hood
FluxNova-3Aura-2Aura-1
as of 2026-07-05
Limitations
- Accuracy can vary by accent and domain compared to some competitors.
- Free tier limited to $200 credit; no perpetual free tier.
- Concurrency limits on lower tiers: STT up to 50 REST, 150 WSS on Pay-as-you-go, up to 225 WSS on Growth.
- Self-hosted and custom models require Enterprise plan.
- API-first design has a learning curve for non-developers.
as of 2026-06-30
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Annual total
Free
Over 12 months
Effective monthly
Free
Billed monthly
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Deepgram tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Pay As You Go
$0/mo ($200 free credit)
Ideal for
Developers and startups exploring Voice AI with no upfront commitment
What this tier adds
Free $200 credit then pay-as-you-go; no minimums, no expiration
Growth
$4K+/year pre-paid credits
Ideal for
Growing applications with predictable usage that want to save up to 20%
What this tier adds
Pre-paid $4K+/year credits with higher concurrency (WSS STU up to 225, TTS up to 60, Voice Agent up to 60)
Enterprise
Contact Sales
Ideal for
Large businesses needing custom concurrency, self-hosting, or custom models
What this tier adds
Contact Sales for custom models, self-hosted deployment, and dedicated support with custom SLAs
Integrations
Amazon ConnectTwilioAsteriskPython SDKNode.js SDKGo SDK.NET SDKJava SDKREST APIWebSocketPipecatLiveKitGoogle Dialogflow CXGenesysAudioCodes
Where the pricing makes sense
The company stage and team size where Deepgram's pricing actually pencils out — and where peers do it cheaper.
Deepgram's Pay As You Go tier with a $200 free credit suits small teams and startups. Growth saves up to 20% for growing apps ($4K+/year). Enterprise offers custom concurrency and self-hosting for large deployments. Compared to AssemblyAI and Google Cloud STT, Deepgram's unified API and lower per-minute rates (e.g., Nova-3 at $0.0048/min streaming) can be cheaper at scale.
Setup time & first value
How long it actually takes to get something useful out of Deepgram — broken out by persona, not the marketing-page minute.
Developers can start with Deepgram's API in minutes: sign up, get a free API key, and use the API Playground. For production, expect a few hours to integrate the Voice Agent API. Self-hosting or custom models (Enterprise) may take weeks.
Switching to or from Deepgram
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
Migrating in
- →From AssemblyAI: Replace your AssemblyAI endpoint with Deepgram's unified Voice Agent API for STT+TTS+LLM—no separate orchestration needed.
- →From Google Cloud Speech-to-Text: Switch to Deepgram's Nova-3 for lower latency and simpler billing—use the REST/WebSocket API.
- →From Rev.ai: Migrate your batch transcription pipeline to Deepgram's Nova-3 batch endpoint with diarization.
Migrating out
- ↗To AssemblyAI: Export transcript data via Deepgram API, then import into AssemblyAI's API.
- ↗To Google Cloud Speech-to-Text: Replace Deepgram calls with Google's API for STT/TTS; note higher latency.
- ↗To Whisper (self-hosted): Use Deepgram's Whisper Cloud endpoint for compatibility, or train custom Whisper models.
Resources & Guides
- Documentationdeepgram.com
Welcome to Deepgram's Docs!
Deepgram developer documentation — APIs, SDKs, and tools for speech-to-text, text-to-speech, voice agents, and audio intelligence.
- Learndeepgram.com
Resources and Tools Created to Inspire
Explore resources and tools created to inspire creativity, perform deep learning, and equip your company with Voice AI best practices.
- Resourcedeepgram.com
Support
Get help with Deepgram — AI-powered support in Slack, community forums, and direct assistance
- Resourcedeepgram.com
Changelog
Helpful link from deepgram.com
- Resourcedeepgram.com
Welcome to the AI Glossary
The Deepgram AI Glossary: Your definitive resource on the world of machine learning, applied deep learning, and the field of Language AI.
- Resourcedeepgram.com
AI Minds The Podcast
Discover how the world’s top companies are building with an AI-first approach on the AI Minds podcast, brought to you by Deepgram. Listen now!
Frequently Asked Questions
Tools that pair well with Deepgram
Common stack mates teams adopt alongside Deepgram, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Alternatives to Deepgram
View allAssemblyAI
Speech-to-text and voice agent APIs for building production-ready voice AI.
FreemiumTry
ElevenLabs
Ultra-realistic AI voice generation, cloning, and conversational agents platform.
FreemiumTry
Whisper Memos
AI voice recorder for iPhone & Apple Watch with email summaries and intelligent routing via Agents.
PaidTry
Used Deepgram? Help shape our editorial sentiment research.