Canary Speech vs Ellipsis Health (2026)
Two vocal biomarker companies that took opposite routes to market. Canary Speech sells the model rather than the application. Its engine extracts roughly 2,590 acoustic and linguistic features from the voice every ten milliseconds, and partners embed it in their own products. That choice shows in what it publishes: model cards naming its certifications with versions, an intended use statement, named compatible record systems, and a documented oversight structure with a threshold, an override and a route to a person. Ellipsis Health sells the finished conversation. Sage is an agentic care manager calling members for health plans and at risk providers, handling enrollment, assessments, surveys and care coordination, powered by what the company calls its Empathy Engine, built on patented vocal biomarker technology and trained on millions of live clinical calls. Its governance has a named process: an ethics committee approving healthcare applications and a clinical team validating every update before release. So the choice is rarely between them. One is bought by a product team with something to build; the other is bought by a plan with members to reach.
- You are building, not buying a workflow: the engine embeds in your product, and the model card tells your team what it does and does not cover.
- The oversight design is published end to end, with a stated threshold, an override and a route to a person.
- Epic, Cerner, Oracle Health and Meditech are named as compatible record systems, and its certifications are named with their versions.
- You need member contact at scale: inbound and outbound calls for enrollment, assessments, surveys and care coordination, aimed at the American payer market.
- The governance process is named twice over, an ethics committee approving healthcare applications and a clinical team validating every product update before deployment.
- The voice work is productised already, so nothing has to be built around it.
This comparison is published by AI Health Index, an independent research platform that compares healthcare AI vendors objectively. Canary Speech and Ellipsis Health are each graded against the same capability taxonomy, from each vendor's own public materials and the regulatory record, under the AI Health Index verification standard. No vendor pays for placement, and no vendor has reviewed this page. How this evidence is graded
Plain facts
| Fact | Canary Speech | Ellipsis Health |
|---|---|---|
| Primary category | Behavioral Health AI | Patient Voice Agents |
| Founded | 2017 | 2017 |
| Headquarters | Provo, Utah, United States | San Francisco, California |
| Website | canaryspeech.com | ellipsishealth.com |
Side by Side
Select any grade to read the note behind it.
Product updates
The latest logged changes from each vendor, newest first, each linked to its source.
Canary Speech
-
EHR / interoperability
Canary Ambient for Physicians is generally available inside Microsoft Dragon Copilot, listed on the Microsoft Marketplace as part of the Dragon Copilot AI Apps and Agents program Microsoft opened on 21 August.
Source
Each vendor in one paragraph
Written to be quoted whole. Each paragraph opens the AI Health Index profile of the vendor: what the product is, who buys it and what it does.
Canary Speech sells the model rather than the application. Its vocal biomarker engine extracts roughly 2,590 acoustic and linguistic features from the human voice every 10 milliseconds and returns behavioral and cognitive indicators from as little as 20 to 45 seconds of natural speech, delivered through an API that other products embed.
Source: AI Health Index, September 2026
Voice AI care management for health plans and at risk providers. The company originated in vocal biomarker research for detecting depression and anxiety from acoustic features, backed by more than ten peer reviewed publications, and that technology is still the engine.
Source: AI Health Index, September 2026
Questions buyers ask
How do Canary Speech and Ellipsis Health work with the EHR?
For Canary Speech, the published model card lists Epic, Cerner and Oracle Health, and Meditech as compatible record systems, which covers the great majority of the American acute market.
With Ellipsis Health, the company names two: a major customer relationship platform's health cloud, where it maintains a listing in that vendor's marketplace, and through a 2026 partnership, a care management suite where it draws on member data and returns outcomes to the platform. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified September 20, 2026. No vendor pays for placement.
How do Canary Speech and Ellipsis Health handle patient data?
In Canary Speech's case, the published model card writes it into the procedure: the clinician obtains informed consent from the patient regarding the recording of their voice before conducting an assessment.
For Ellipsis Health, a privacy policy, a terms of use and, separately, a consumer health data privacy policy are all published, the last being the notice that state consumer health data statutes require. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified September 20, 2026. No vendor pays for placement.
Which settings and specialties do Canary Speech and Ellipsis Health cover?
With Canary Speech, the company now sells by segment: healthcare providers, payers, clinical call centers, contact centers and clinical trials, with condition lines for behavioral, cognitive, neurological, safety and wellness. February 2026 added consumer caregiving, embedding passive analysis into video calls between older adults and their families, its first deployment outside clinical and research environments.
In Ellipsis Health's case, the customer boundary is explicit: health plans and at risk provider organizations, not consumers and not clinical settings. The workflow boundary is equally clear and enumerated on the product page across three groups. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified September 20, 2026. No vendor pays for placement.
What evidence is published for Canary Speech and Ellipsis Health?
For Canary Speech, the 2026 paper in the Proceedings of Artificial Intelligence in Medicine reports unweighted average recall of 0.70 for depression and 0.68 for anxiety, and a combined assessment reaching sensitivity of 0.76 with specificity of 0.65 on a remote test set. Performance was then examined on an independent in clinic dataset collected by tablet, where sensitivity was 0.68 and specificity 0.80.
With Ellipsis Health, a partnership embeds the agent into a major care management suite drawing on member data and returning outcomes to the platform, a second places it inside a large customer relationship platform's health cloud, and a third, announced with a leading artificial intelligence hardware company, concerns more natural conversation. From the AI Health Index, drawn from each vendor's own published materials and the regulatory record, verified September 20, 2026. No vendor pays for placement.
Related comparisons
Other published head to head assessments involving these vendors or their closest peers. The full set for this category is on the Remote Monitoring & Chronic Care page.
Both assert that their voice models work across populations and neither publishes per group results, which is the standing gap in this technology: speech varies by accent, dialect, age and first language, and medication alters the very features these models read. Ask both for per group figures. Canary's published performance is vendor authored and honest about its numbers, with commercial outcome evidence still missing, and its compliance position has no contracting statement behind it.
Ellipsis publishes no escalation criterion for a voice agent making outbound calls to recently discharged members, and its consent question is the sharper one: the vendor states that vocal cue detection runs in production, so a member who called about a prescription may have an inference drawn from their voice and should be told where it goes. Neither publishes a price.