Current recording support list. Historical training across 30+ languages does not expand this product-support contract.
Published training data
Provider describes training on hundreds of thousands of speech hours across 30+ languages, followed by tuning on thousands of hours of de-identified medical conversations.Source for Published training data (opens in a new tab)
Phrases are assigned to speakers; no fixed maximum speaker count or raw diarization API contract is published.
Speaker attribution
Automatically predicts speaker roles in encounters, including patient, clinician, caregiver and interpreter; no manual speaker configuration is required.Source for Speaker attribution (opens in a new tab)
Integrated clinical product through the provider or health-system deployment. A separate inference API for this family/component was not established.
Regions
Serving and storage location must be verified under the relevant customer agreement.
Endpoint pricing has not been collected. Check the provider for applicable rates.
Data & enterprise terms
The August 27, 2026 public privacy policy explicitly excludes customer clinical content from its general scope. That content, including PHI, is governed by customer agreements and BAAs. Exact model processing region, clinical recording retention and training-use terms are not established by the public website policy. No blanket zero-retention or no-training guarantee is inferred.
Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.
01 / INITIAL INPUTProcess & retain
The system processes eligible input and retains reusable state.
02 / LATER REQUESTCheck available state
Reuse depends on matching input, session state and the cache policy.
03 / CACHE HITReuse & generate
Reusing state can reduce repeated work. Any charges follow the billing rules below.
Abridge clinical platform
Not verified
No reviewed inference-cache contract. Retained recordings, prior-note context and enrolled voiceprints are not prompt caching.
How it works
Not collected
No reviewed inference-cache contract. Retained recordings, prior-note context and enrolled voiceprints are not prompt caching.
Minimum reusable input
Not collected
No reviewed inference-cache contract. Retained recordings, prior-note context and enrolled voiceprints are not prompt caching.
Lifetime & refresh
Not collected
No reviewed inference-cache contract. Retained recordings, prior-note context and enrolled voiceprints are not prompt caching.
How to enable it
Not collected
No reviewed inference-cache contract. Retained recordings, prior-note context and enrolled voiceprints are not prompt caching.
Usage fields
Not collected
No reviewed inference-cache contract. Retained recordings, prior-note context and enrolled voiceprints are not prompt caching.
Billing rules
Not collected
No reviewed inference-cache contract. Retained recordings, prior-note context and enrolled voiceprints are not prompt caching.
Limits & data boundaries
Not collected
No reviewed inference-cache contract. Retained recordings, prior-note context and enrolled voiceprints are not prompt caching.
No verified cache rate for this endpoint. An empty rate is not a zero charge.
Caching reviewed 16 Oct 2026
What the measurements say
No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.
Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.
The record behind the model
Why it is included
Provider explicitly identifies an in-house proprietary trained model family. The exact checkpoint, base ancestry and downloadable weight release are not exposed in the reviewed sources.
