AssemblyAI
“Can a developer-first Voice AI infrastructure company become the default speech and language understanding layer across every AI-native product?”
Founded2017
HQSan Francisco, CA, USA
Latest roundSeries C, December 2023
IndustryInfrastructure / AI / Speech-to-Text
The story
Founded in 2017 as a developer-focused speech-to-text API company backed by Y Combinator. Expanded from basic transcription into a full Voice AI infrastructure platform, adding speech understanding, sentiment analysis, PII redaction, and entity detection. As of 2024–2025, pivoted toward real-time and agentic use cases with the launch of Streaming Speech-to-Text and Voice Agent API, positioning AssemblyAI as the infrastructure layer for AI-native voice products rather than a standalone transcription service.
Last 12 months
Product timeline
2017
Founded and backed by Y Combinator, launching initial speech-to-text API infrastructure.· pivot
2023
Raised $50M Series C led by Accel to build superhuman Voice AI models.· banking
2024
Launched Universal-3 Pro, the company's flagship high-accuracy multilingual speech model.· pivot
2025
Launched Voice Agent API enabling real-time voice agent construction with streaming diarization and 99+ language support.· pivot
The stack
Banking / BaaS
Accounting gap: none