> AssemblyAI
AssemblyAI sells speech-to-text, speech understanding and voice agent APIs with pay-as-you-go hourly pricing and no credit card required to start. The free tier is published as up to 185 hours of pre-recorded transcription and up to 333 hours of streaming transcription. Published rates are $0.15 per hour for Universal-2 (99 languages) and $0.21 per hour for Universal-3.5 Pro (18 languages with native code switching), while realtime runs at $0.15 per hour for Universal-Streaming (English) and multilingual, or $0.45 per hour for Universal-3.5 Pro Realtime. The Voice Agent API is $4.50 per hour ($0.075 per minute) with every feature included, and the Dictation API is $0.62 per hour.
// pros
- Free tier is stated in hours, not credits: up to 185 hours of pre-recorded transcription and up to 333 hours of streaming transcription with no credit card required.
- Streaming concurrency scales automatically at no extra cost: the free plan allows 5 new streaming connections per minute, pay-as-you-go starts at 100 sessions per minute and rises by 10% whenever 70% or more of the limit is used, with no ceiling and custom starting limits above 100 at no additional charge.
- Add-on features are individually priced and mostly small, for example speaker diarization at $0.02 per hour, profanity filtering at $0.01 per hour, PII redaction of the transcript at $0.08 per hour and keyterm prompting at $0.05 per hour.
- The $4.50 per hour Voice Agent API has no per-layer add-ons, concurrency fees or per-agent subscriptions, so one line on the bill covers the whole request.
- The Dictation API returns finished text from a single spoken utterance in about 134 ms at p50 with clips up to 120 seconds, across 19 languages, at $0.62 per hour.
// cons
- Free-tier streaming starts at only 5 new connections per minute, so bursty production traffic needs the pay-as-you-go plan even though concurrency then scales automatically.
- The most accurate models are the most expensive: Universal-3.5 Pro Realtime is $0.45 per hour versus $0.15 per hour for Universal-Streaming, and the 18-language Pro model is priced above the 99-language Universal-2 at $0.15 per hour.
- Several capabilities are tier-restricted rather than merely priced: prompt-based prompting is not supported on streaming, and voice isolation is streaming-only at $0.10 per hour.
- Understanding features are billed per hour of audio in addition to transcription, so sentiment ($0.02 per hour), entity detection ($0.08 per hour), topic detection ($0.15 per hour) and translation ($0.06 per hour) stack onto the base rate.
Start with the free tier's 185 pre-recorded hours to benchmark Universal-2 against Universal-3.5 Pro on your own audio, then choose per hour rather than per feature: base transcription at $0.15 per hour plus diarization at $0.02 per hour remains cheaper than moving the whole workload onto the $0.21 per hour Pro model if you do not need its 18-language code switching. For live voice agents, model the $4.50 per hour all-in Voice Agent API instead of assembling streaming plus LLM plus orchestration separately.
written by Mike Fleming, Uprouter editorial · verified against official sources September 16, 2026
Value taken from the provider’s own published free-tier terms, expressed in USD.
note: Free tier is stated as up to 185 hours of pre-recorded transcription and up to 333 hours of streaming transcription with no credit card required; valued at the lowest published pre-recorded rate (Universal-2 at $0.15 per hour) this is about $27.75, an estimate because AssemblyAI publishes hours rather than a USD credit (assemblyai.com/pricing, checked 2026-09-16).
| Plan | Type | Monthly | Input /1M | Output /1M | Note |
|---|---|---|---|---|---|
| Universal-3.5 Pro (STT) | payg | — | Free | Free | $0.21 per hour of audio. |
| LLM Gateway (GPT-5 Nano) | payg | — | $0.050 | $0.400 | — |
| Universal-2 (pay-as-you-go) | payg | — | — | — | $0.15 per hour of audio transcribed |
| Universal-3.5 Pro (pay-as-you-go) | payg | — | — | — | $0.21 per hour of audio transcribed; custom/enterprise rates on request |
| Free trial | free | — | — | — | Start free, then pay-as-you-go; no commitments (free minutes amount not restated on current pricing page) |
Pricing normalized from public sources — always verify with the provider.
// model prices
| Model | Input /1M | Output /1M | Context |
|---|---|---|---|
| Universal-3.5 Pro | — | — | — |
| Universal-2 | — | — | — |
| Ubuntu | — | — | — |
| Universal-2 | — | — | — |
| Universal (Streaming) | — | — | — |
$ Does AssemblyAI have a free tier?
Yes — AssemblyAI offers a free tier that we value at roughly $27.75 in free credits. Always confirm current limits on AssemblyAI's official pricing page — free tiers change without notice.
$ How much does AssemblyAI cost per 1M tokens?
AssemblyAI does not publish simple per-1M-token pricing (it may bill per request, per output, or via a subscription plan). Check their official pricing page for exact figures.
$ Is AssemblyAI safe to use?
Uprouter rates AssemblyAI as low risk. Low to medium risk. Transcription is priced per hour of audio with small, individually published add-on rates, which makes spend predictable, and the free tier is a genuine allocation of up to 185 pre-recorded hours rather than a time-limited trial. The main cost risks are stacking: understanding features and the premium realtime model are billed on top of the base hour rate, so an unmeasured pipeline can multiply the headline price several times. Concurrency is the operational risk, since free-plan streaming starts at 5 new connections per minute and production bursts require the paid tier. Risk ratings are editorial, evidence-based, and never influenced by affiliate relationships.
Raw community sentiment, shown unblended. Votes never feed into objective data fields like price, risk, or status — and never into sort order.
No published reviews yet — be the first.
// more_informationprice history · live status · risk & confidence · code snippets
price_history
No recorded changes yet.
live_status
current: unknownNo probes recorded yet.
Uprouter runs its own probes; best-effort, not a guarantee. Probe results depend on our network location and cadence — for production, always use the provider’s own status page.
risk_&_confidence
Low to medium risk. Transcription is priced per hour of audio with small, individually published add-on rates, which makes spend predictable, and the free tier is a genuine allocation of up to 185 pre-recorded hours rather than a time-limited trial. The main cost risks are stacking: understanding features and the premium realtime model are billed on top of the base hour rate, so an unmeasured pipeline can multiply the headline price several times. Concurrency is the operational risk, since free-plan streaming starts at 5 new connections per minute and production bursts require the paid tier.
Confidence reflects how much verifiable evidence (official docs, probe history, community reports) backs this entry. Risk is our editorial assessment of operator reliability and terms — not financial advice.
code_snippets
# AssemblyAI's own API — get a key from their dashboard
curl https://assemblyai.com/chat/completions \
-H "Authorization: Bearer THEIR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "assemblyai/universal-3.5-pro",
"messages": [{ "role": "user", "content": "Hello" }]
}'Direct to the provider — Uprouter receives nothing; verify their current pricing before purchasing.
Ranked by neutral similarity signals only — live status, free-tier shape and free-credit proximity. Commercial relationships never influence this list.