Rev is a US-based speech-to-text company founded in 2010 by former oDesk employees Jason Chicola and Josh Bardsley. Headquartered in Austin, Texas, the company has grown into one of the largest transcription and captioning providers in the world, blending an on-demand workforce of human transcriptionists with proprietary AI speech recognition models offered under the Rev.ai brand.
The consumer-facing product at rev.com lets customers order automated AI transcripts, 99% accuracy human-reviewed transcripts, captions, and subtitles directly from a browser or mobile app. The platform is widely used by legal professionals, court reporters, journalists, podcasters, and academic researchers who need defensible, court-admissible transcripts. Subscription plans range from a free tier with limited AI minutes to Basic, Pro, and Enterprise tiers, alongside pay-per-minute options at roughly $0.25 per AI minute and $1.99 per human minute.
Rev.ai is the developer-facing arm of the business and exposes the same underlying ASR engine through REST and streaming APIs, with SDKs for building voice-powered applications. In 2024-2025 Rev rolled out Reverb, a next-generation foundational speech model priced as low as $0.20 per audio hour, alongside features for sentiment analysis, topic extraction, entity detection, language identification, and speaker diarization across 36+ languages.
On the funding side, Rev has raised approximately $51.5M across its early venture rounds — the most prominent being a 2013 Series A led by Globespan Capital Partners. Unlike many AI startups, the company crossed $100M in annual recurring revenue and reported profitability before scaling further, giving it an unusually self-sustaining capital profile. As of 2025 the company reports thousands of contracted Revvers across six continents.
Rev differentiates itself by offering both ends of the transcription spectrum under one roof: cheap, fast AI for high-volume needs and human-verified accuracy for regulated or high-stakes use cases. That dual model is particularly attractive to law firms, depositions, and media organizations that cannot tolerate the error rates typical of pure-AI competitors.