San Francisco-based Rime announced it has secured $24 million in a Series A financing round led by M13 Ventures, with participation from Twilio Ventures, Corazon Capital, Unusual Ventures and existing backers. The company was founded in 2022 by Lily Clifford (a former Stanford PhD student), Brooke Larson (ex-Amazon Alexa engineer) and Ares Geovanos (Stanford engineer).
What distinguishes Rime
Rime is positioning its voice-AI models to handle enterprise calls across sales, marketing and customer support, a crowded space that already includes players such as ElevenLabs, Deepgram, infrastructure providers like Vapi, Retell and LiveKit, and dedicated support companies such as Decagon and Sierra. Rather than relying on scraped web audio, Rime built a recording studio in San Francisco to collect conversational data for training its models.
The company focuses on tuning models to accurately pronounce brand names and industry-specific terminology. It uses a phoneme-based architecture to adapt to different pronunciations, aiming to reduce the need for customers to retrain models for their particular vocabulary.
Technical shift: fewer models, more speech-to-speech
Rime originally operated separate pipelines for speech-to-text, text-to-speech and a large language model. The startup is now shifting toward developing improved speech-to-speech models to reduce latency, improve turn-taking in conversations and better handle background noise. This approach also intends to lower orchestration overhead by relying on more integrated models rather than managing many discrete components.
Customers and enterprise traction
Rime says it serves customers across food service, healthcare, airlines and fintech, citing clients such as Mayo Clinic, Dialpad, Upstart and Asurion. The company claims its training data and model positioning lead to longer call durations, which has helped it win enterprise contracts.
Market realities and limitations
Founder Lily Clifford observes that despite advances in voice AI, many enterprises still favor legacy IVR (interactive voice response) systems because current voice AI experiences do not yet match IVR effectiveness. Clifford noted that large language models have made it easier to build functional voice applications, but the interaction experience still feels limited: “Talking with a voice AI agent is not the most compelling experience for the end user. It’s kinda like a new IVR, but with a better voice.”
Team growth and hires
Rime has a team of 35 employees and intends to use the new funding to hire more people in model development, engineering and partnerships. The company recently appointed Rafael Valle as Chief Scientist; Valle previously worked on audio understanding at Meta Superintelligence Labs and on applied deep learning audio research at NVIDIA.
M13’s Morgan Blumberg told TechCrunch that while some companies have shifted toward orchestration and the application layer, substantial technical work remains, and he sees Rime’s focus on low-latency, reliable models in regulated environments as a differentiator.
Prior fundraising
Rime raised $5.5 million in a seed round last May. As part of the Series A, Morgan Blumberg will join the company’s board.
