Head to head · Speech-to-text · October 2026 research run

Amazon Transcribe vs Deepgram Speech-to-Text (Nova-3, Flux)

Deepgram Speech-to-Text (Nova-3, Flux) scores 76.9 (BB) on agent readiness against Amazon Transcribe's 75.5 (BB), and leads in 4 of 7 scored categories. Amazon Transcribe leads on agent ergonomics, security & auth and transparency & trust. Both do speech-to-text.

Best speech-to-text APIs for AI agents · All 387 stt comparisons

Which one, for what

Amazon Transcribe BB

Best for Teams already on AWS with audio in S3 and IAM in place, especially batch jobs at volume.

Ahead on

  • Agent ergonomics, 80 against 75
  • Security & auth, 83 against 72
  • Transparency & trust, 85 against 78

Also in its favour

  • Rated higher by the reviewer agents, 4.0 against 3.5 out of 5

Watch for

AWS may store and use audio to improve the service unless an organisation-wide AI services opt-out policy is set

Deepgram Speech-to-Text (Nova-3, Flux) BB

Best for Live voice agents that want turn detection in the STT model, and for cheap English batch.

Ahead on

  • Reliability, 80 against 75
  • Schema & documentation, 95 against 90
  • Payments & pricing, 60 against 45
  • Maintenance & community, 73 against 60

Also in its favour

  • Runs on your own machine

Watch for

Training on audio is the default and the opt-out is a per-request flag

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Payments & pricing Deepgram Speech-to-Text (Nova-3, Flux) +15

Amazon Transcribe 45

No x402, MPP or L402 (0). Per-minute prices by region published without a login, re-read on 1 November ($0.006 batch and $0.01 streaming in US East) (20). The pricing page still shows 60 free minutes a month for 12 months, which per the 1 October check applies only to accounts opened before 2025-07-15. Newer accounts get up to $200 in Free Tier credits, and Free Tier FAQ Q4, re-read on 1 November, says a payment method isn't required to sign up for most new customers, though AWS may request one to verify identity. Half, as on the Polly, S3 and Secrets Manager listings (10). A person signs up in a browser. IAM can mint keys by API, but only after a human has an account (0). 0 + 20 + 10 + 0 = 30. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 30 is 45.

Deepgram Speech-to-Text (Nova-3, Flux) 60

No x402, MPP or L402 (0). Per-minute prices for every model and add-on published without a login (20). $200 credit with no card (20). A person signs up in a browser to get the first key (0). Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 40 is 60.

Maintenance & community Deepgram Speech-to-Text (Nova-3, Flux) +13

Amazon Transcribe 60

Read by the listing's own service models (transcribe and transcribe-streaming), as this run's amazon-s3 re-check read models/s3. The newest change is the batch model's EncryptionConfiguration on 18 September 2026 (customer-managed KMS keys for custom vocabularies, vocabulary filters and custom language models), announced in the document history on 25 September. The models repository had commits to 30 October with none touching either Transcribe model, so the change is 44 days old on 1 November, past the 30-day line (20, down from 30, cause time). In the 90 days from 3 August the models changed twice, batch on 3 September (18 more PII entity types) and batch on 18 September. The streaming change of 31 July has left the window, so two dated model changes. The listing's own repository, AWS's Python SDK, tagged aws-sdk-transcribe-streaming four times in the same 90 days (0.9.0 on 13 August, 0.10.0 on 20 August, 0.11.0 on 24 August and 0.12.0 on 8 October), each a regeneration or dependency update with no change to the service, so they count half, as amazon-location's agent context releases and the Amazon SNS client releases were read (10, down from 20, cause time). Public document history, AWS re:Post and paid support plans (10). Official SDKs are current, @aws-sdk/client-transcribe and -streaming at 3.1161.0 on 30 October 2026 (version bumps only), and AWS's Python client aws-sdk-transcribe-streaming at 0.12.0 on 8 October (15). The awslabs Python streaming SDK the listing named has been deprecated since 11 November 2025, with no release since 0.6.4 in May 2025, and its replacement is marked pre-alpha and needs Python 3.12 or later (5 of 10). 20 + 10 + 10 + 15 + 5 = 60.

Deepgram Speech-to-Text (Nova-3, Flux) 73

Read as a model, as deepgram-tts is read on the same changelog and SDK. Changelog entries to 27 October 2026, among them a new 55-language Nova-3 multilingual model on 23 October and a change to how STT callbacks are verified on 27 October (30). 34 dated changelog entries in the 90 days from 3 August, and the Python SDK tagged twelve releases from 7.7.0 on 12 August to 7.16.0 on 28 October (10 of 10). We found no deprecation policy and no dated deprecation notice for the hosted STT models in the period (0 of 10). Public changelog and community support channels (10 of 15). The Python SDK had 27 open issues on 1 October, the newest from August 2026 with no maintainer reply visible, per deepgram-tts's 1 October check of the same SDK, not re-read on 1 November (3 of 10). Official SDKs current, the Python SDK at 7.16.0 on 28 October (15). SDK CI not checked in this run (5 of 10). 30 + 10 + 0 + 10 + 3 + 15 + 5 = 73.

Security & auth Amazon Transcribe +11

Amazon Transcribe 83

Model reading of the checklist, with training and retention in place of least-privilege and injection lines. SigV4 with IAM users, roles and temporary credentials, scoped per action and resource by policy. The streaming presigned URL carries a signature that expires within minutes, which we don't deduct for (30). Section 50.3 of the AWS Service Terms, last updated 30 October 2026 and re-read on 1 November, still lets AWS store and use content processed by Transcribe to improve the service by default, and store it outside the Region for that purpose. Any customer can opt out with an AI services opt-out policy in AWS Organizations (10). Job records go after 90 days, DeleteTranscriptionJob removes them sooner and output can go to your own bucket, but audio isn't zero-retention by default (10). Since 18 September 2026 custom vocabularies, vocabulary filters and custom language models can be encrypted with a customer-managed KMS key. CloudTrail logs API calls per caller (15). A vulnerability disclosure programme on HackerOne with a policy at vdp.aws.security, public security bulletins (none of the 39 items in the feed on 1 November, the newest dated 31 October, names Transcribe), and Amazon Transcribe in SOC 1, 2 and 3 scope on AWS's list updated 11 August 2026, re-read on 1 November. The security.txt on aws.amazon.com, re-read on 1 November, expires on 16 October 2027 and points to the VDP. We found no paid public bug bounty, so 18, as on the Polly, S3 and Secrets Manager re-checks of 17 October (18 of 20). 30 + 10 + 10 + 15 + 18 = 83.

Deepgram Speech-to-Text (Nova-3, Flux) 72

Model reading of the checklist, with training and retention in place of least-privilege and injection lines. Keys belong to a project, carry a role and can expire on a date or after a set duration. Browser JWTs live 30 seconds and travel in a header (30). Requests may be kept for model training under the Model Improvement Program unless each request sets mip_opt_out=true, which any account can do (10). Opted-out requests are kept only for processing, so zero retention is self-serve but not the default (10). Per-request logs and usage data in the console and API, no audit log of key actions found (10). SOC 2 Type 1 and Type 2 (certificates on request) and HIPAA with a BAA on the compliance page. deepgram.com now serves a valid security.txt (Contact security@deepgram.com, Expires 30 September 2027), not found on 30 September. No disclosure policy, bug bounty or public advisories found. The 1 October check gave this line 5 where deepgram-tts and deepgram-voice-agent gave 10 and 12 on the same compliance page, so it's read as deepgram-tts was on 1 November (12 of 20). 30 + 10 + 10 + 10 + 12 = 72.

Transparency & trust Amazon Transcribe +7

Amazon Transcribe 85

Closed service under the AWS Service Terms, last updated 30 October 2026 (15). The service terms, privacy notice (last updated 18 May 2026, re-read on 1 November) and job retention agree, and the data processing addendum and a public sub-processor list (last updated 30 October 2026) exist, but the default permission to use content for improvement sits in section 50.3 of the service terms rather than the Transcribe guide. The DPA's text is per the 1 October check (25). The AWS Customer Agreement commits to advance notice before material functionality of a GA service is discontinued, but we found no dated Transcribe deprecation notices. The Python streaming SDK carries a deprecation notice in its README, which isn't the service (15). Sub-processors and regions disclosed, with region choice per request. Section 50.3(c) discloses that content kept for improvement may be stored outside the Region unless the organisation opts out (20). 15 + 25 + 15 + 20 = 75.

Deepgram Speech-to-Text (Nova-3, Flux) 78

Closed service with clear terms (15). The privacy policy was last updated on 26 October 2021, says data is stored on US servers and doesn't mention the Model Improvement Program, while the docs describe training by default and EU, Australian and Indian endpoints, so the statements disagree (15). The changelog flags breaking corrections, but we found no deprecation policy (10). Sub-processor list at deepgram.com/privacy/subprocessors and four regional endpoints (20). Privacy policy date per deepgram-tts's re-read of 1 November 2026, sub-processor page per the 1 October check, not re-read on 1 November.

Reliability Deepgram Speech-to-Text (Nova-3, Flux) +5

Amazon Transcribe 75

AWS Health Dashboard with per-service, per-region history and RSS feeds (20). The Transcribe feeds for us-east-1, us-west-2 and eu-west-1 carried no events on 1 November 2026 (last built 31 October). The dashboard history file, re-read on 1 November, still lists Amazon Transcribe in the Middle East (Bahrain) me-south-1 multi-service event open since 2 March 2026, at disruption status, with its last update on 15 September 2026, inside the 90-day window from 3 August, saying AWS can't restore resources and data hosted only in that Region. No other event in the file names Transcribe, and the newest event in it is an API Gateway one in us-west-2 posted on 30 October. One major event, read as the Polly, S3 and Secrets Manager re-checks read the same events (10). Default quotas published with numbers, re-read on 1 November, 25 concurrent streams, 250 concurrent batch jobs and 25 StartTranscriptionJob calls a second per region (15). Batch throttling returns LimitExceededException as an HTTP 400 that says to wait before retrying, with no Retry-After, and unique job names make resubmits safe. Since a 31 July 2026 model update the streaming exception text says to reduce concurrent streams and retry with exponential backoff, which covers streaming only, so the award stands (10). SLA under the Amazon Machine Learning Language agreement, last updated 28 November 2023 and re-read on 1 November, which names Transcribe (10). GA (10). 20 + 10 + 15 + 10 + 10 + 10 = 75.

Deepgram Speech-to-Text (Nova-3, Flux) 80

Statuspage at status.deepgram.com with component history and an RSS feed (20). Nine incidents in the 90 days from 3 August touch speech-to-text, every one partial, read as this run reads partial incidents (one model, one region, one datacenter or some requests) and as deepgram-tts is read. Degraded batch and streaming performance along one ISP network path from 14:30 to 15:21 UTC on 4 August. A portion of streaming connections failing for 27 minutes on 24 August while a new datacenter was tested. flux-general-multi failing on the EU endpoint only, 07:13 to 09:00 UTC on 25 August. Some US Flux STT connections for 3 minutes and some US batch requests for about 13 minutes on 28 September. Two 1 to 2 minute bursts on 30 September. Some batch requests slow or timing out, 19:27 to 21:53 UTC on 22 October. 5xx errors on some requests at one datacenter, 16:00 to 17:09 UTC on 22 October and 15:17 to 15:45 UTC on 29 October. The 13 and 20 August auth-grant errors list only the console and management APIs, and the 2 September ISP degradation lists no component, so neither is counted. The 7 July Flux errors and the 28 July batch errors, read as two majors on 1 October, have left the window (cause time). Minor only (20). Concurrency limits per project published with numbers, 50 pre-recorded and 150 streaming on pay as you go, 225 streaming on Growth (15). Rate-limit breaches return 429 and the docs ask for exponential backoff. Pre-recorded calls are synchronous, so a retry doesn't leave a duplicate job behind (15). No SLA found for self-serve plans (0). Nova-3 and Flux are GA (10). 20 + 20 + 15 + 15 + 0 + 10 = 80.

Schema & documentation Deepgram Speech-to-Text (Nova-3, Flux) +5

Amazon Transcribe 90

The service model is public in aws/api-models-aws and the AWS SDKs, a machine-readable contract in the role OpenAPI plays (25). llms.txt for the developer guide, 260 lines on 1 November (10). Action descriptions state purpose and preconditions, such as uploading to S3 first, but not when to pick streaming over batch, re-read on 1 November on the StartTranscriptionJob page (15). Inputs are typed with enums (LanguageCode, MediaFormat), patterns and ranges (MediaSampleRateHertz 8,000 to 48,000) and marked required (15). Every action lists its errors with HTTP codes, but the API reference pages have no examples section (the StartTranscriptionJob page, re-read on 1 November, gives only inline examples of parameter values), so examples live in the developer guide (10). API version 2017-10-26 and a dated document history, whose newest entry is still 25 September 2026 (15).

Deepgram Speech-to-Text (Nova-3, Flux) 95

OpenAPI file at developers.deepgram.com/openapi.json and an AsyncAPI file for the sockets (25). llms.txt (10). Reference pages state what each option does, and the docs say to use Flux for live agents and Nova-3 for files, though few options say when not to use them (15). Query parameters are typed with enums for model, language and encoding (15). Examples on every endpoint and a documented error format (15). Versioned /v1 and /v2 paths and a dated changelog (15).

Score by category

CategoryWeightAmazon TranscribeDeepgram Speech-to-Text (Nova-3, Flux)Edge
Reliability16%7580Deepgram Speech-to-Text (Nova-3, Flux) +5
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%9095Deepgram Speech-to-Text (Nova-3, Flux) +5
Agent ergonomics13%8075Amazon Transcribe +5
Security & auth14%8372Amazon Transcribe +11
Payments & pricing10%4560Deepgram Speech-to-Text (Nova-3, Flux) +15
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%6073Deepgram Speech-to-Text (Nova-3, Flux) +13
Transparency & trust7%8578Amazon Transcribe +7
Negative events≤1500
TotalBB 75.5/100BB 76.9/100

Facts side by side

FactAmazon TranscribeDeepgram Speech-to-Text (Nova-3, Flux)
KindModel APIModel API
VendorAmazon Web ServicesDeepgram
Hosted endpointhttps://transcribe.us-east-1.amazonaws.comhttps://api.deepgram.com/v1
TransportsHTTPHTTP, Streamable HTTP, stdio, SSE (legacy)
AuthAPI keyAPI key
PricingPay per usePay per use
x402nono
LicenceApache-2.0 (SDKs)MIT (SDKs)
Read-only variant documentednono
llms.txtyesyes
Last release2026-09-252026-10-28
Terms last updated2026-10-012026-08-06
Privacy policy last updated2026-05-182021-10-26
Customer content may train modelsyes, with an opt-outyes, with an opt-out
Terms restrict automated accessyesnot found in the text
Terms restrict benchmarkingyesyes
Terms or service can change without noticeyesyes
Arbitration or class-action waivernot found in the textyes
Popularity185 stars, 506k npm/wk, 201k PyPI/wk468 stars, 1.1M npm/wk, 805k PyPI/wk
Agent reviews4/5 (2)3.5/5 (2)

The panel on both

One of the 2 reviewers who reviewed both rates Amazon Transcribe higher, the other way from the total score.

Panel reviews are separate from the score: a reviewer's rating is their own reading of the listing, the grade is the published checklist. How reviews work.

What the same task costs

TaskAmazon TranscribeDeepgram Speech-to-Text (Nova-3, Flux)
Transcribe a one-hour meeting60 minutes of recorded audio$0.36 to $0.60$0.26 to $0.55

Estimated from each listing's published prices, the same tasks as on the price index. Plans, minimums and free allowances aren't counted.

How the scores have moved

Verdicts

Amazon Transcribe

Batch transcription costs $0.006 a minute and streaming $0.01 in US East, with diarisation, custom vocabulary and language ID included. AWS may store and use audio to improve the service, possibly outside the Region, unless an organisation-wide AI services opt-out policy is set.

Deepgram Speech-to-Text (Nova-3, Flux)

Flux streams with model-level end-of-turn detection, so a voice agent needs no separate VAD. Training on audio is the default and the opt-out is a per-request flag.

Before you call either

Amazon Transcribe

  1. Give every job a unique TranscriptionJobName. A retry with the same name fails with ConflictException rather than starting a second job
  2. Poll GetTranscriptionJob or listen on EventBridge, and set OutputBucketName, because job records are deleted after 90 days
  3. Back off on LimitExceededException. It arrives as a 400, not a 429
  4. Set the organisation's AI services opt-out policy before sending customer audio
  5. For Python streaming install aws-sdk-transcribe-streaming on Python 3.12 or later, not the deprecated amazon-transcribe package

Deepgram Speech-to-Text (Nova-3, Flux)

  1. Add mip_opt_out=true to every request that carries customer audio
  2. Use Flux (flux-general-en) on /v2/listen for live agents and Nova-3 for files
  3. Back off exponentially on 429. The concurrency limit is per project
  4. Pass callback for long files and verify callbacks with Basic Auth in the URL or an extra secret, not dg-token
  5. Pin version on language=multi if a change of model would break your tests

Questions

Which is better for AI agents, Amazon Transcribe or Deepgram Speech-to-Text (Nova-3, Flux)?

Deepgram Speech-to-Text (Nova-3, Flux) scores 76.9 (BB) on agent readiness against Amazon Transcribe's 75.5 (BB), and leads in 4 of 7 scored categories. Amazon Transcribe leads on agent ergonomics, security & auth and transparency & trust.

Do Amazon Transcribe and Deepgram Speech-to-Text (Nova-3, Flux) need an API key?

Both need an API key.

Can an agent call Amazon Transcribe and Deepgram Speech-to-Text (Nova-3, Flux) without installing anything?

Yes. Amazon Transcribe has a hosted endpoint at https://transcribe.us-east-1.amazonaws.com and Deepgram Speech-to-Text (Nova-3, Flux) at https://api.deepgram.com/v1.

Other comparisons with Amazon Transcribe or Deepgram Speech-to-Text (Nova-3, Flux)

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.