Rev AI is a developer-first speech-to-text API, a separate product from Rev.com's human transcription service. Whether it suits you depends on wanting transcription plus add-on APIs, like translation, summarization and sentiment analysis, all billed by usage from one provider rather than a flat plan.
What is Rev AI?
Rev AI is rev.ai's speech-to-text API, a separate product from Rev.com's human transcription service. It runs as a cloud API with a web dashboard for account and token management; no self-hosted or on-premises deployment is offered, only regional data processing in the United States and the European Union. Rev AI is headquartered in the United States.
What the Rev AI API does
Rev AI transcribes audio asynchronously and in real time, through a dedicated Streaming Speech-to-Text API. Its documentation names 58+ supported languages for the asynchronous API and 9+ for the streaming API. Reverb ASR is the core model, with Reverb Transcription priced at $0.20 an hour for English and Reverb Foreign Language Transcription at $0.30 an hour, covering Spanish, French, Chinese, Portuguese and more than 50 further languages; Whisper Large is offered as an additional model option.
Speaker diarization runs by default on every audio file processed through the async API, separating and labelling who spoke when. The JSON transcript output includes a timestamp for every word. A dedicated Custom Vocabulary API lets you supply domain-specific phrases to improve recognition. Rev AI does not offer text-to-speech; it is a speech-to-text product only.
Working with Rev AI
Rev AI adds a set of add-on APIs alongside transcription: Forced Alignment, Language Identification, Language Translation in Standard and Premium models, Sentiment Analysis, Summarization in Standard and Premium models, and Topic Extraction, each billed separately. Transcripts come back in Rev AI's own JSON schema. Enterprise accounts can enforce Single Sign-On using SAML 2.0 across a domain or team.
Rev AI pricing: what each plan gives
Rev AI has no flat monthly or annual plan: Pay As You Go is pure usage billing, and Enterprise pricing is volume-based and quoted directly. On Pay As You Go, Reverb Transcription costs $0.20 per hour of English audio and Reverb Foreign Language Transcription costs $0.30 per hour; Whisper Large costs $0.005 per minute. The add-on APIs are billed separately: Forced Alignment and Language Identification cost $0.003 a minute, Standard-model translation and summarization cost $0.002 a minute against $0.025 a minute for the Premium models, and Sentiment Analysis and Topic Extraction cost $0.0008 per 10 words.
New accounts get free credit equivalent to 5 hours of Reverb ASR, usable across all of Rev AI's products.
Privacy and support
Rev AI never uses your content to train external AI models and does not sell customer data or allow it to be used to train external LLMs. It is SOC 2 Type II audited, and its services are designed to help customers meet their own GDPR obligations. Enterprise customers get a HIPAA-compliant transcription tier and can enforce SAML 2.0 single sign-on.
Support runs through email and a documentation help centre; no live chat or phone line is offered.
Who Rev AI suits
Rev AI suits development teams who want speech-to-text billed strictly by usage, plus optional add-ons like translation, summarization and sentiment analysis from the same provider, rather than assembling several separate tools. Its no-training stance on customer content and SOC 2 audit also suit teams handling sensitive audio.
It suits less well anyone who wants text-to-speech in the same product, since Rev AI is transcription only, or anyone who needs to run the model inside their own infrastructure, since there is no self-hosted option, only regional cloud data residency.
Frequently Asked Questions
How much does Rev AI cost?
Rev AI has no subscription; Pay As You Go is billed by usage. Reverb Transcription costs $0.20 an hour of English audio, and new accounts start with free credit equivalent to 5 hours of Reverb ASR.
How many languages does Rev AI transcribe?
Rev AI's documentation states 58+ languages for the asynchronous Speech-to-Text API and 9+ languages for the Streaming Speech-to-Text API.
Can Rev AI be self-hosted?
No. Rev AI runs only as a cloud API, with regional data processing available in the United States and the European Union, not as a self-hosted or on-premises deployment.
Does Rev AI train on the audio I send it?
No. Rev AI never uses customer content to train external AI models and does not allow that data to be used to train external LLMs.
