AI Tools
Deepgram
Deepgram : prix, capacités, mises à jour et alternatives, lus sur les pages de l’éditeur.

En anglais
Key facts
- Price
- Free plan · paid plans: see pricing Read on deepgram.com
- Free plan
- Yes Read on deepgram.com
- Platforms
- Web · CLI · API Read on deepgram.com
- API
- Yes
- Open source
- Not presented as open source
- Last update
- 16 Nov 2026 Read on developers.deepgram.com
- Prices checked
- 26 Oct 2026 Read on deepgram.com
Deepgram is a developer platform for adding speech and voice features to applications. Its APIs handle real-time or recorded audio transcription, text-to-speech, audio analysis, and voice-agent interactions.
Developers can use cloud or regional endpoints, or self-host in an environment that meets the requirements. Usage is billed by product; the free offer is $200 in credits, while Growth uses prepaid annual credits and Enterprise pricing is custom. Self-hosting is available on Enterprise.
- Category
- Audio & Music
- Read from the maker’s pages
- 2026-11-18
Company
Features
- Transcribe real-time audio streams and prerecorded files
- Generate speech from text
- Build voice agents with speech-to-text, a managed language model, and text-to-speech
- Analyze audio for summaries, topics, sentiment, and intent
- Detect speakers and redact sensitive information from transcripts
- Support language detection and code-switching
- Run through regional endpoints or self-host on Enterprise
Use cases
- Transcribe interviews, meetings, podcasts, and support calls
- Add live captions to events and applications
- Build voice agents that handle turn-taking and interruptions
- Analyze customer calls for topics, sentiment, and intent
- Generate spoken responses for voice applications
Pricing
Free plan · paid plans: see pricing
- Prices checked
- 2026-10-26
- Read
- by web search
Pay As You Go
free
- Speech-to-Text
- Up to 50 REST / 150 WSS / 5 Whisper Cloud
- Text-to-Speech
- Up to 45 REST + WSS
- Voice Agent API
- Up to 45 WSS
- Free $200 Credit then pay-as-you-go
- No minimums. No expiration. No credit card required.
- All endpoints in public models
Growth
from $4000/yr
- $4K+ / year
- Save up to 20% with pre-paid credits for the year
- Credits redeemed against actual usage
- All endpoints in public models
- Higher concurrency: Speech-to-Text up to 50 REST / 225 WSS / 5 Whisper Cloud; TTS up to 60; Voice Agent up to 60
- Community & Discord support, Standard Uptime SLA
Enterprise
price on request
- For businesses with large volumes, data or deployment requirements, or support needs
- Contact Sales
Capabilities
- Text to speech — “Converts AI responses into natural-sounding speech.” source
- Speech to text — “Real-time transcription, production accuracy” source
- Many languages — “Conversational speech recognition for real-time voice agents that handle multiple languages within a single conversation” source
- API — “Deepgram unifies speech-to-text, text-to-speech, and LLM orchestration into a single API, reducing complexity, latency, and cost.” source
Latest updates
- Nova-3 Adds 36 Languages: Abkhazian, Adyghe, Akan, Albanian, Amharic, Azerbaijani, Basque, Bodo, Central Kurdish, Divehi, Eastern Mari, Gal…
We’ve added 36 Nova-3 monolingual languages for batch and streaming workloads.
- Improved Nova-3 Models for Catalan and Japanese
We’ve released improved Nova-3 monolingual models for Catalan and Japanese for batch and streaming workloads.
- Improved Nova-3 Models for Arabic, Armenian, Bengali, Bulgarian, Danish, English (Australia, New Zealand), Greek, Gujarati, Hebrew, Nepali,…
We’ve released improved Nova-3 monolingual models for 13 existing languages for batch and streaming workloads.
- Deepgram Self-Hosted November 2026 Release (261110)
Container images include the self-hosted API, engine, license proxy, and billing.
- Verify Callback Requests with Basic Auth or Extra Metadata
The dg-token header is no longer supported; switch to Basic Auth or use a secret passed with Extra Metadata for Speech-to-Text.
Where it works
Web · CLI · API
Alternatives to Deepgram
- AssemblyAI
Speech AI APIs for transcription, audio analysis, dictation, and text-to-speech.
- Ursa
Speechmatics' Ursa is a speech-recognition model family that transcribes audio into text for batch and real-time applications.
- ElevenLabs
ElevenLabs makes AI tools for generating and editing speech, audio, images and video, transcribing speech, and building voice and chat agents.
- Saaras
Sarvam’s Saaras converts speech into text, translation, transliteration, or code-mixed output across Indian languages and English.
- Fish Audio
Fish Audio turns text into expressive speech and offers voice cloning, transcription, and voice-agent tools.
- Rime Mist
Rime Mist is a text-to-speech model for generating streaming speech in voice agents and other real-time applications.
Questions about Deepgram
Is Deepgram free?
Yes. The maker’s pages state that Deepgram can be used for free. Read on the maker’s pages on 18 Nov 2026.
How much does Deepgram cost?
The plans on the maker’s pricing page: Pay As You Go (free), Growth (from $4000/yr) and Enterprise (price on request). These prices were read by web search, not on the maker’s own page: check its pricing page before relying on them. Prices checked on 26 Oct 2026.
Does Deepgram have an API?
Yes. The maker’s pages state that Deepgram has an API. Read on the maker’s pages on 18 Nov 2026.
Which platforms does Deepgram run on?
Deepgram runs on Web, CLI and API, as its maker’s pages state. Read on the maker’s pages on 18 Nov 2026.
Is Deepgram open source?
The maker’s pages do not present Deepgram as open source. Read on the maker’s pages on 18 Nov 2026.
Is Deepgram still available?
Yes, as of 18 Nov 2026: its maker’s pages were read that day and carried no notice that Deepgram has ended.
What changed recently in Deepgram?
The latest entry in its maker’s changelog: 16 Nov 2026, “Nova-3 Adds 36 Languages: Abkhazian, Adyghe, Akan, Albanian, Amharic, Azerbaijani, Basque, Bodo, Central Kurdish, Divehi, Eastern Mari, Gal…”. Before it: 12 Nov 2026, “Improved Nova-3 Models for Catalan and Japanese”. Changelog checked on 17 Nov 2026.
Sources
- https://deepgram.com/pricing?gspk=Y29sb3JtYW5nb2xlZTk4MzU&gsxid=p0f4YaT7l2GgIi&ps_partner_key=Y29sb3JtYW5nb2xlZTk4MzU&ps_xid=p0f4YaT7l2GgIi — Official pricing page read today; lists plans, prices, usage rates, and self-hosting availability.
- https://deepgram.com/product/speech-to-text — Official product page describes transcription capabilities, language handling, deployment options, and use cases.
- https://deepgram.com/ — Official homepage identifies Deepgram's voice AI platform and its speech, text-to-speech, and voice-agent offerings.
- https://developers.deepgram.com/changelog — Official changelog lists the November 17, 2026 Nova-3 model improvement.
Not confirmed
- Latest version
- Deepgram offers multiple models and the pages read do not identify one overall current product version.