Subtle Computing is a California-based startup developing end-to-end voice isolation technology that lets computers reliably understand users in challenging acoustic settings such as loud cafés or open offices. Instead of training a single generic model, the company trains device-specific models that preserve each device’s acoustic characteristics, delivering an order-of-magnitude performance improvement over standard solutions. Its core voice-isolation model can run entirely on-device, occupies only a few megabytes, and operates with roughly 100 ms latency; a complementary model converts the cleaned audio into highly accurate text transcripts. Subtle Computing has already been selected for Qualcomm’s Voice & Music Extension Program, ensuring compatibility with OEM products, and it is piloting its software with an unnamed consumer hardware brand and an automotive manufacturer. Looking ahead, the team plans to launch its own consumer product that combines hardware and software sometime next year. The company was founded by Stanford alumni Tyler Chen, David Harrison, Savannah Cofer, and Jackie Yang and has secured $6 million in seed financing to date.

Total raised

$6M

Funding rounds

1

Latest round

Seed

Latest activity

Nov 2025