Deepgram
Deepgram
★ #16 in Audio GenerationProduct status: Status not recorded
No lifecycle review is recorded for this listing. Its inclusion in the directory does not confirm current availability; check the vendor before relying on it.
No status check date recorded
Speech recognition, speech synthesis and integrated voice-agent APIs with model-specific usage billing. No lifecycle review is recorded for this listing. Its inclusion in the directory does not confirm current availability; check the vendor before relying on it.
Product preview
Decision notes
Official-document research
Based on the cited documentation. We did not use the product, purchase a plan, or run quality, speed or reliability tests in this review.
Consider it for
- Developers managing audio rights, provider selection and usage budgets
Choose another approach if
- A flat unlimited plan, assumed no-training defaults or standalone resale
Pricing and conditions
Documented terms
USD plus applicable taxes. PAYG starts with USD200 credit; Growth is USD4,000+ annually prepaid against usage, not a monthly plan. PAYG examples: Nova-3 monolingual recorded USD0.0043/audio minute; streaming currently promotional 0.0048 (regular 0.0077), end date unconfirmed. Aura-2 USD0.030/1,000 characters; Voice Agent Standard USD0.075/minute or BYO LLM+TTS 0.050, external provider charges separate. Add-ons may cost extra. PAYG auto-load and Growth annual auto-renewal apply; credits/refunds depend on plan/terms. Silence, channel multipliers, rounding, overage details and regional totals not verified.
Pricing source read:
Pricing or terms sourceCode, model and service licensing
Service integration is permitted in an application with material independent functionality; standalone resale and competitive model training/analysis are restricted. You keep rights you possess in inputs/outputs, not model weights, with no uniqueness or third-party clearance guarantee. Audio/personality rights and synthetic-content notices remain your responsibility. Terms permit perpetual content use including training/testing by default. The MIP opt-out is per request; docs say opted-out request data is retained only as needed to process it. This does not cover external LLM practices or guarantee zero account/log retention; business data retention follows agreement. No source/model license was established.
Basis and limits of the recorded score
Original 4.5/5 retained; no original worksheet or hands-on basis available. Not re-scored.
Overview
Deepgram provides hosted transcription, synthesis and voice-agent APIs. Voice Agent combines speech recognition, an LLM and speech output over WebSocket, with tool calls and provider configuration. Flux adds turn detection; Nova transcription and feature/language coverage differ. This review does not establish latency, accuracy or savings against alternatives. Pricing and entitlement are separate questions: confirm which models, languages, diarisation and add-ons a plan includes, and whether usage is billed by audio hour or by token before comparing vendors.
Sources
The factual claims in this review were read from the pages below. Each date is when an editor last opened the source, not when the product changed.
- Vendor pricing page Checked 2026-10-15
- Vendor documentation Checked 2026-10-15
- Vendor documentation Checked 2026-10-15
- Vendor documentation Checked 2026-10-15
- Vendor documentation Checked 2026-10-15
- Vendor documentation Checked 2026-10-15
Pros / Cons
Pros
- Integrated voice-agent pipeline
- Explicit model/endpoint pricing and request-level training opt-out
Cons
- Add-ons and external providers affect total cost
- Training opt-out must be configured; it is not a blanket data guarantee
Who it's for
Transcription workflows with reviewed speaker labels
Voice applications with controlled tool access
Bottom line
Estimate each model, add-on and external-provider meter, then configure data choices before sending audio.
Key Features
Tags
Frequently asked questions
Is Deepgram currently available?
No lifecycle review is recorded for this listing. Its inclusion in the directory does not confirm current availability; check the vendor before relying on it.
What is Deepgram?
Deepgram is a Audio Generation tool developed by Deepgram. Speech recognition, speech synthesis and integrated voice-agent APIs with model-specific usage billing.
What costs and terms should I check for Deepgram?
USD plus applicable taxes. PAYG starts with USD200 credit; Growth is USD4,000+ annually prepaid against usage, not a monthly plan. PAYG examples: Nova-3 monolingual recorded USD0.0043/audio minute; streaming currently promotional 0.0048 (regular 0.0077), end date unconfirmed. Aura-2 USD0.030/1,000 characters; Voice Agent Standard USD0.075/minute or BYO LLM+TTS 0.050, external provider charges separate. Add-ons may cost extra. PAYG auto-load and Growth annual auto-renewal apply; credits/refunds depend on plan/terms. Silence, channel multipliers, rounding, overage details and regional totals not verified.
Is Deepgram open source?
Service integration is permitted in an application with material independent functionality; standalone resale and competitive model training/analysis are restricted. You keep rights you possess in inputs/outputs, not model weights, with no uniqueness or third-party clearance guarantee. Audio/personality rights and synthetic-content notices remain your responsibility. Terms permit perpetual content use including training/testing by default. The MIP opt-out is per request; docs say opted-out request data is retained only as needed to process it. This does not cover external LLM practices or guarantee zero account/log retention; business data retention follows agreement. No source/model license was established.
What are the best alternatives to Deepgram?
Notable Audio Generation alternatives include ElevenLabs, Suno, Udio. Browse all 51 tools in our Audio Generation category for a full overview.
Was this listing useful?
Basic Info
- Pricing
- Pay-as-you-go
USD plus applicable taxes. PAYG starts with USD200 credit; Growth is USD4,000+ annually prepaid against usage, not a monthly plan. PAYG examples: Nova-3 monolingual recorded USD0.0043/audio minute; streaming currently promotional 0.0048 (regular 0.0077), end date unconfirmed. Aura-2 USD0.030/1,000 characters; Voice Agent Standard USD0.075/minute or BYO LLM+TTS 0.050, external provider charges separate. Add-ons may cost extra. PAYG auto-load and Growth annual auto-renewal apply; credits/refunds depend on plan/terms. Silence, channel multipliers, rounding, overage details and regional totals not verified.
- Editorial score
- ★ 4.5 / 5 How we score
- Platforms
- API
- Last verified
- Not recorded
