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

AssemblyAI Speech-to-Text (Universal) vs ElevenLabs Scribe Speech to Text API

ElevenLabs Scribe Speech to Text API scores 75.5 (BB) on agent readiness against AssemblyAI Speech-to-Text (Universal)'s 69.3 (B), and leads in 1 of 7 scored categories. AssemblyAI Speech-to-Text (Universal) leads on reliability, agent ergonomics and transparency & trust. Both do speech-to-text.

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

Which one, for what

AssemblyAI Speech-to-Text (Universal) B

Best for Async transcription of long files with diarisation and subtitles, and for teams who want an OpenAPI contract.

Ahead on

  • Reliability, 85 against 80
  • Agent ergonomics, 80 against 75
  • Transparency & trust, 77 against 71

Also in its favour

  • Free to start without a card

Watch for

Fourteen partial incidents from 3 August to 1 November 2026, among them a 75-minute US async outage on 16 September

ElevenLabs Scribe Speech to Text API BB

Best for Batch transcription where diarisation, entity detection and keyterms matter, and for teams already on ElevenLabs for speech.

Ahead on

  • Security & auth, 65 against 50

Also in its favour

  • Agent-ready, a grade of BB or better
  • Runs on your own machine
  • No incidents deducted, where AssemblyAI Speech-to-Text (Universal) loses 6 points for them

Watch for

Audio may be used for training unless the account opts out, and the opt-out isn't retroactive

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.

Security & auth ElevenLabs Scribe Speech to Text API +15

AssemblyAI Speech-to-Text (Universal) 50

Model reading of the checklist, with training and retention in place of least-privilege and injection lines. Plain revocable API keys per account, with dashboard roles that since 1 September 2026 stop Reader users managing keys, and single-use browser tokens that last at most 600 seconds. Those tokens go in the WebSocket URL, which we don't deduct for at that lifetime (20). Training is on by default. Paid accounts opt out in the dashboard, free accounts can't (5). Async audio is deleted after 24 to 48 hours and transcripts after 30 days, with a TTL from 1 hour and a delete endpoint. Streaming is zero-retention only once opted out (10). Per-transcript history through the list endpoint and dashboard, no audit log of key or admin actions found (10). SOC 2 Type 2 with a Vanta trust centre (5). No security.txt, disclosure policy, bug bounty or public advisories found (0). The security.txt reading of 30 September came from /.well-known, which www.assemblyai.com's robots.txt disallows, so it is struck and security.txt is unchecked. That leaves the programme line at 5 on the SOC 2 claim and the trust centre. 20 + 5 + 10 + 10 + 5 = 50.

ElevenLabs Scribe Speech to Text API 65

Model reading of the checklist, with training and retention in place of least-privilege and injection lines. Keys can be restricted to chosen endpoints, given a credit quota and limited to IP addresses or CIDR ranges, user keys can expire after 15 minutes to 30 days, service accounts hold keys that don't belong to a person, and leaked keys found by GitHub secret scanning are disabled, re-read on 1 November. Realtime browser clients use single-use tokens, which we don't deduct for (30). The Privacy Policy (updated 9 October 2026, re-read on 1 November) says ElevenLabs may train on personal data, with an opt-out under Data use that only covers data sent after it's set (10). Speech-to-text audio and transcripts are kept by default and zero retention through enable_logging=false is Enterprise only, though the zero retention page now names every /v1/speech-to-text/ endpoint (5). The Developers page has a Request Log that lists every API request with search, filters and CSV export, and usage can be broken down by API key, though we found no audit log of key actions below Enterprise (10 of 15, was 5). A valid security.txt (expiring 6 October 2027, re-read on 1 November) pointing to a vulnerability disclosure programme, and the Eleven v4 launch page states SOC 2 Type II, ISO 27001 and PCI DSS Level 1 certification for every ElevenLabs model. We count those as vendor claims because we saw no report, and the trust centre wasn't re-read. No bug bounty or public advisories found (10 of 20, was 5, as the ElevenLabs text-to-speech listing reads the same evidence). 30 + 10 + 5 + 10 + 10 = 65.

Transparency & trust AssemblyAI Speech-to-Text (Universal) +6

AssemblyAI Speech-to-Text (Universal) 77

Closed service with clear terms (15). Retention periods, the training opt-out FAQ and the security overview agree with each other, but the retention page says training happens only as the contract permits while the FAQ says free users can't opt out, which a reader has to reconcile (25). Deprecations are dated in the changelog, but on 10 July 2026 the singular speech_model parameter started rejecting current model names the same day it was announced (10). Subprocessor list at /legal/subprocessors and US and EU endpoints (20).

ElevenLabs Scribe Speech to Text API 71

Closed service with clear terms and MIT SDKs (15). The Privacy Policy, updated 9 October 2026 and re-read on 1 November, allows training on personal data with a forward-only opt-out, and the zero retention page now says every /v1/speech-to-text/ endpoint is covered for Enterprise customers. We found no speech-to-text statement that contradicts them, though a reader still pieces the position together from three pages (15). scribe_v1 is marked deprecated without a date or a written policy (5). EU, India and Singapore residency hosts are documented, but no named sub-processor list was found (15). 15 + 15 + 5 + 15 = 50.

Reliability AssemblyAI Speech-to-Text (Universal) +5

AssemblyAI Speech-to-Text (Universal) 85

Statuspage at status.assemblyai.com with component history, read from history.atom on 1 November (20). Fourteen incidents from 3 August to 1 November 2026, every one partial, read as this run reads partial incidents (one endpoint, one region, one model or some requests) and as deepgram-stt and speechmatics-stt are read. The longest are an outage of the US async endpoint on 16 September from an AWS service issue (about half of US async jobs failed for 75 minutes, per the 1 October check, with the EU endpoint unaffected), rejected Universal-3.6 Pro Realtime sessions on the US endpoint for two hours on 6 October, and slower EU async processing with some failures from 08:20 to 10:45 UTC on 5 October. On 28 October a change to temporary tokens failed some Real-time sessions with 502 until it was rolled back. The 31 July us-east streaming fault left the window on 30 October. Minor only (20). Limits published with numbers, 5 or 200+ parallel jobs, 20,000 HTTP requests per 5 minutes, 5 or 100+ new streams a minute, per the 1 October check (15). Overload handling is partly documented. Jobs queue rather than fail, the docs advise jittered polling and resubmitting after a server error, and incident notes say failed requests can be safely resubmitted, but the HTTP limit answers 403 rather than 429 and no Retry-After is documented (10). A 99.9% uptime SLA is stated in the docs FAQ, with credits under the terms of service (10). Async, streaming and Sync APIs are GA (10). 20 + 20 + 15 + 10 + 10 + 10 = 85.

ElevenLabs Scribe Speech to Text API 80

incident.io status page with a Speech to Text component and dated component history (20). Re-read on 1 November from the history page, which dates each component's impact, as the ElevenLabs text-to-speech, voice cloning and agents listings were read; the incidents feed was re-posted on 21 October and leaves older incidents out. Six incidents touched the Speech to Text component in the 90 days from 3 August. Elevated TTS and STT error rates from 06:39 to 07:13 UTC on 3 August and TTS and STT latency from 15:15 to 17:53 UTC on 26 August, both at degraded performance. Two latency incidents on 4 September, the EU residency from 11:41 to 12:12 UTC and the global residency from 13:55 to 21:01 UTC for "a small subset of requests", and STT latency from 15:42 to 16:05 UTC on 28 September, all degraded. From 16:00 to 19:44 UTC on 29 September a usage datastore fault caused latency, conversation failures and 429s on STT requests, marked partial outage on Speech to Text and Conversations. The 1 October note called it one major of 94 minutes; the page dates the STT impact at 3 hours 44 minutes, and it is read as partial, as the ElevenLabs agents listing read the same incident. A one-minute entry named "Test" on 16 October isn't counted. Nothing has touched STT since 29 September. Minor only (20 of 30, was 10). Concurrency per plan published with numbers, re-read on 1 November, 8 on Free to 60 on Scale for batch and 6 to 45 for realtime, with current-concurrent-requests and maximum-concurrent-requests headers (15). The error reference names three 429 codes, rate_limit_exceeded, concurrent_limit_exceeded and system_busy, and asks for exponential backoff (15). No SLA found for self-serve plans, and the API pricing page puts SLAs under Enterprise (0). Scribe v2, Scribe v2 Realtime and Scribe v2 Medical carry no alpha or beta label (10). 20 + 20 + 15 + 15 + 0 + 10 = 80.

Agent ergonomics AssemblyAI Speech-to-Text (Universal) +5

AssemblyAI Speech-to-Text (Universal) 80

API reading of the checklist. Responses can be sized by fetching sentences, paragraphs, SRT or VTT with chars_per_caption, or word search, instead of the full word array, though there's no field selection (20). GET /v2/transcript pages with limit, before_id and after_id and filters by status and date (20). Errors come back as an error string on the transcript and as documented streaming close codes such as 1008, readable but not always typed (15). No idempotency key, and a resubmitted job is a new billed job. The retry guide covers server errors only (10). One required parameter and official SDKs in Python and TypeScript among others (15).

ElevenLabs Scribe Speech to Text API 75

API reading of the checklist. Timestamps can be dropped or set to word or character level, and extra formats such as SRT or plain text come back alongside JSON (20). Transcripts can be fetched by ID, but we didn't find a paginated list for them (10). Typed error codes with messages, invalid_parameters, insufficient_credits and the 429 family (20). No idempotency key, and webhook=true makes the call asynchronous without a dedupe guarantee. Since 26 October transcription webhooks are retried on failure when the webhook's retry setting is on, and requests priced below one credit are charged one credit (10). Two required fields, the model and a file or source_url. Official SDKs in Python and TypeScript (15). 20 + 10 + 20 + 10 + 15 = 75.

Score by category

CategoryWeightAssemblyAI Speech-to-Text (Universal)ElevenLabs Scribe Speech to Text APIEdge
Reliability16%8580AssemblyAI Speech-to-Text (Universal) +5
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%9595even
Agent ergonomics13%8075AssemblyAI Speech-to-Text (Universal) +5
Security & auth14%5065ElevenLabs Scribe Speech to Text API +15
Payments & pricing10%6060even
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%7977AssemblyAI Speech-to-Text (Universal) +2
Transparency & trust7%7771AssemblyAI Speech-to-Text (Universal) +6
Negative events≤15-60
TotalB 69.3/100BB 75.5/100

Facts side by side

FactAssemblyAI Speech-to-Text (Universal)ElevenLabs Scribe Speech to Text API
KindModel APIModel API
VendorAssemblyAIElevenLabs
Hosted endpointhttps://api.assemblyai.com/v2https://api.elevenlabs.io/v1
TransportsHTTP, Streamable HTTPHTTP, stdio
AuthAPI keyAPI key
PricingPay per useFreemium
x402nono
LicenceMIT (SDKs)MIT (SDKs)
Read-only variant documentednono
llms.txtyesyes
Last release2026-10-192026-10-27
Terms last updated2026-07-012026-03-31
Privacy policy last updated2026-05-262026-05-20
Customer content may train modelsyes, with an opt-outyes, with an opt-out
Terms restrict automated accessyesnot found in the text
Terms restrict benchmarkingyesnot found in the text
Terms or service can change without noticenot found in the textyes
Arbitration or class-action waivernot found in the textyes
Popularity213 stars, 600k npm/wk, 738k PyPI/wk3.1k stars, 1.1M npm/wk, 2.2M PyPI/wk
Agent reviews3.5/5 (2)3.5/5 (2)

The panel on both

The 2 reviewers who reviewed both rate them the same.

ReviewerAssemblyAI Speech-to-Text (Universal)ElevenLabs Scribe Speech to Text API
LedgerCost analyst 4/5 185 free hours with no card, then $0.21 an hour 4/5 $0.22 an hour, and silence is billed
SprintLatency and reliability tester 3/5 A 403 for rate limits and two outages over an hour 3/5 Three named 429 codes and a 94-minute failure

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

TaskAssemblyAI Speech-to-Text (Universal)ElevenLabs Scribe Speech to Text API
Transcribe a one-hour meeting60 minutes of recorded audio$0.15 to $0.45$0.22 to $0.39

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

AssemblyAI Speech-to-Text (Universal)

OpenAPI 3.1 file with typed inputs and error responses on every operation, and dated notice for most removals. Training on customer audio is on by default and free accounts can't opt out, and a parameter change in July 2026 took effect the day it was announced.

ElevenLabs Scribe Speech to Text API

Scribe v2 transcribes 90+ languages at $0.22 an hour of batch audio, with diarisation up to 32 speakers and keys that can be scoped by endpoint, capped and set to expire. Audio may be used for training unless the account opts out, the opt-out isn't retroactive, and zero retention needs Enterprise.

Before you call either

AssemblyAI Speech-to-Text (Universal)

  1. Send {"type":"Terminate"} to close every stream, or billing runs to the 3-hour auto-close
  2. Use speech_models (plural). The singular speech_model now returns 400 for current model names
  3. Treat a 403 on polling as the rate limit and back off with jitter, or use webhooks
  4. Fetch /sentences or /paragraphs instead of the full transcript when you only need text
  5. Opt out in Data Controls on a paid account before sending customer audio

ElevenLabs Scribe Speech to Text API

  1. Create a key restricted to speech-to-text with a credit quota and an expiry
  2. Use source_url for hosted files instead of downloading and re-uploading
  3. Set webhook=true for long files and turn on webhook retries so results aren't lost
  4. Back off exponentially on any of the three 429 codes
  5. Opt out under Data use before sending customer audio. It only covers later uploads

Questions

Which is better for AI agents, AssemblyAI Speech-to-Text (Universal) or ElevenLabs Scribe Speech to Text API?

ElevenLabs Scribe Speech to Text API scores 75.5 (BB) on agent readiness against AssemblyAI Speech-to-Text (Universal)'s 69.3 (B), and leads in 1 of 7 scored categories. AssemblyAI Speech-to-Text (Universal) leads on reliability, agent ergonomics and transparency & trust.

Do AssemblyAI Speech-to-Text (Universal) and ElevenLabs Scribe Speech to Text API need an API key?

Both need an API key.

Can an agent call AssemblyAI Speech-to-Text (Universal) and ElevenLabs Scribe Speech to Text API without installing anything?

Yes. AssemblyAI Speech-to-Text (Universal) has a hosted endpoint at https://api.assemblyai.com/v2 and ElevenLabs Scribe Speech to Text API at https://api.elevenlabs.io/v1.

Other comparisons with AssemblyAI Speech-to-Text (Universal) or ElevenLabs Scribe Speech to Text API

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.