Heidi Health chooses Argmax as the on-device inference engine for their popular AI medical scribe app that supports 2 million consults every week and returns 8 million hours to clinicians every year.
On-device deployment with Argmax enabled Heidi Health to cut the end-of-session latency to less than 1 second (p95) independent of the session length and internet connectivity, upgrading the user experience where clinicians can start reviewing the finalized transcript right after the patient visit ends. This improvement comes without any compromise to Heidi's note accuracy that users have come to trust.
Heidi also improved their cost structure from usage-based (variable) cloud API pricing to license-based (fixed) SDK pricing. This helps sustain their highly generous free-tier with unlimited transcripts as they continue to scale in usage.
What The CEO Has to Say
We are incredibly excited to partner with Argmax to bring fast, accurate and compliant transcription to Heidi devices around the world. This is pushing the bleeding edge of software and hardware integration to deliver a fully offline AI care partner for our clinicians and patients.
- Dr. Thomas Kelly, Co-founder and CEO at Heidi Health
Argmax continues to run real-time speech-to-text even on the Lock Screen, unlike other engines that are shut off by iOS when the app is backgrounded. This feature was a requirement for Heidi's ambient listening experience where clinicians' put their phones aside shortly after starting the recording session and the phones are automatically screen-locked after being idle for a few minutes.
