OpenEvidence
Also known as: DeepConsult, OpenEvidence Visits
Clinical AI app for verified clinicians: evidence-grounded answers from its own model family (Osler, Sackett, Snow), plus Visits documentation, Coding Intelligence and a HIPAA-secure Doctor Dialer; free for verified clinicians.
OpenEvidence is a clinical AI platform for verified US clinicians on the web and iOS and Android. Its own model family answers clinical questions grounded in peer reviewed literature and guidelines at three depths. Osler answers in about five seconds for the encounter, Sackett in about thirty seconds, and Snow in about five minutes with a full investigation of the literature, with Darwin in research preview. Around the answer engine sit three tools. Visits records and transcribes patient interactions into notes.
Coding Intelligence adds ICD-10 diagnoses, E/M level recommendations with MDM rationale and CPT suggestions to finished notes. The Doctor Dialer handles HIPAA secure calls, messages, faxes and voicemail, and can turn a call into a visit note. Health systems such as Mount Sinai bring OpenEvidence into the Epic workflow, and UTMB and Penn Medicine are named EHR partners. The production models are free for verified clinicians.
Vendor details
Canonical URL
https://www.openevidence.com
Category
Healthcare agent
Funding status
$700M+ total at $12B valuation; $250M Series D Jan 2026 (Thrive, DST) after $200M Series C at $6B Oct 2025 (GV); backers include Sequoia, Nvidia, Kleiner Perkins, Blackstone, Mayo Clinic
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Clinicians at Mount Sinai reach OpenEvidence inside their Epic workflow, and UTMB and Penn Medicine are named EHR integration partners. A partnership with Tandem links evidence based decisions to prescription generation and prior authorization submission. No write back to the EHR and no connector list is published.
In practice
Mid visit, you need the current treatment for a patient whose case doesn't fit the usual guideline. You ask OpenEvidence and get a cited answer in about five seconds, with a grade showing how strong the evidence behind it is.
You treat cancer patients and want the NCCN algorithm without leaving the conversation. When you ask, the most relevant NCCN algorithm appears in context beside peer reviewed evidence and a link back to NCCN.
You record the visit and the note is drafted for you, and when you finish it ICD-10 diagnoses, the E/M level with its MDM rationale and CPT suggestions are already there for you to review.
Sources & related URLs
Research sources
Agentic Index coverage score
3.0 / 14 capabilities · 21%
| Integrations & Tool Calling | Partial |
|---|---|
|
Clinicians at Mount Sinai ask questions inside their Epic workflow, and EHR integration partnerships with UTMB and Penn Medicine are named. A partnership with Tandem connects evidence based decisions with prescription generation and prior authorization submission, with Tandem automating prior authorizations and routing each prescription to the preferred pharmacy. OpenEvidence does not say whether it sends anything into those systems itself, or writes back to the EHR. SourceOpenEvidence, globenewswire.com/news-release/2026/03/31/3265707 and businesswire.com/news/home/20260402254260/enread 2026-10-06 |
|
| Workflow Orchestration | Partial |
|
Each answer runs inside a fixed product flow, from Osler's answer in about five seconds to Snow's full investigation of the literature in about five minutes. Visits turns a recorded visit into a note that Coding Intelligence then codes automatically the moment it is finished. OpenEvidence names no branching or flow the health system configures. SourceOpenEvidence, businesswire.com/news/home/20260903878517/en and prnewswire.com/news-releases/openevidence-launches-coding-intelligence-to-help-physicians-capture-every-dollar-theyve-earned-302726278.htmlread 2026-10-06 |
|
| Knowledge Grounding & RAG | Partial |
|
Answers are grounded in peer reviewed literature and clinical guidelines. NCCN oncology guidelines surface in context alongside citations and a link back to NCCN, and Cochrane systematic reviews and the American Academy of Otolaryngology Head and Neck Surgery Foundation's guidelines come in through partnerships. Snow runs a full investigation of the literature before answering. EvidenceGrade grades and shows in real time the quality of the evidence cited in each answer, building on the GRADE framework behind Cochrane, the WHO and most major guidelines. The corpus is OpenEvidence's own curated literature, and OpenEvidence names no way for an institution to add its own knowledge. SourceOpenEvidence, businesswire.com/news/home/20260427952447/en and prnewswire.com/news-releases/openevidence-launches-evidencegrade-empowering-physicians-to-see-the-strength-of-cited-evidence-beneath-each-ai-answer-302822750.html; businesswire.com/news/home/20260903878517/en/introducing-the-openevidence-model-familyread 2026-10-06 |
|
| Human Oversight & Guardrails | Partial |
|
The physician keeps the decision while the platform advises and drafts. Visits drafts the note, and Coding Intelligence is applied automatically at the end of every visit to surface ICD-10 diagnoses, E/M levels with the full MDM rationale and CPT suggestions. OpenEvidence names no approval step before codes or notes are committed. SourceOpenEvidence, prnewswire.com/news-releases/openevidence-launches-coding-intelligence-to-help-physicians-capture-every-dollar-theyve-earned-302726278.htmlread 2026-10-06 |
|
| Security, Identity & Governance | Not documented |
|
OpenEvidence calls its communications HIPAA compliant and offers the Doctor Dialer for HIPAA secure calls. It names no attestation and no access model. The production models are free to clinicians once they are verified. SourceOpenEvidence, openevidence.com and businesswire.com/news/home/20260903878517/enread 2026-10-06 |
|
| Observability & Auditability | Not documented |
|
Answers show the citations they draw on, so a clinician can see where an answer came from. OpenEvidence names no logs, history or admin reporting for a health system to review. SourceOpenEvidence, openevidence.com/sitemap.xml and prnewswire.com/news-releases/openevidence-launches-evidencegrade-empowering-physicians-to-see-the-strength-of-cited-evidence-beneath-each-ai-answer-302822750.htmlread 2026-10-06 |
|
| Memory & State Persistence | Not documented |
|
No memory of a clinician or a patient that carries from one question to the next is published. SourceOpenEvidence, openevidence.com/sitemap.xmlread 2026-10-06 |
|
| Deployment & Data Residency | Not documented |
|
OpenEvidence runs on openevidence.com and in its iOS and Android apps. It says nothing about hosting, regions or customer environments. SourceOpenEvidence, businesswire.com/news/home/20260903878517/en and openevidence.comread 2026-10-06 |
|
| Prebuilt Agents, Templates & Packs | Partial |
|
One clinician app carries several features. Evidence search runs across three model tiers with EvidenceGrade beneath each answer, Visits handles documentation with Coding Intelligence inside it, and the Doctor Dialer covers calls, messages, faxes and voicemail. Coding Intelligence depends on Visits and the rest are modes of one app, so none stands alone as a separate product. The production models are free to verified clinicians. SourceOpenEvidence, businesswire.com/news/home/20260903878517/en and prnewswire.com/news-releases/openevidence-launches-evidencegrade-empowering-physicians-to-see-the-strength-of-cited-evidence-beneath-each-ai-answer-302822750.htmlread 2026-10-06 |
|
| Triggers & Channel Coverage | Partial |
|
A clinician starts the work by asking a question, recording a visit or placing a call through the Doctor Dialer on the web or in the iOS and Android apps. Coding Intelligence follows automatically when a visit note is finished. No schedule or outside event starts work on its own. SourceOpenEvidence, businesswire.com/news/home/20260903878517/en and prnewswire.com/news-releases/openevidence-launches-coding-intelligence-to-help-physicians-capture-every-dollar-theyve-earned-302726278.htmlread 2026-10-06 |
|
| Model Flexibility & Routing | Not documented |
|
Clinicians pick among OpenEvidence's own models by speed and depth. Osler answers in about five seconds for the pace of the patient encounter, Sackett runs a deeper search in about thirty seconds and Snow takes about five minutes. Darwin is open by application only to partners such as the National Organization for Rare Disorders, research collaborators and accredited AI researchers. Every option is OpenEvidence's own model. SourceOpenEvidence, businesswire.com/news/home/20260903878517/enread 2026-10-06 |
|
| APIs, SDKs & MCP Extensibility | Not documented |
|
No API, SDK or MCP server is published. The models are reached only through openevidence.com, the iOS and Android apps and EHR partners. SourceOpenEvidence, businesswire.com/news/home/20260903878517/enread 2026-10-06 |
|
| Testing, Debugging & Optimization | Not documented |
|
Customers get no evaluation tooling of their own. OpenEvidence reports scores for its own models on MedQA, MedXpertQA, HealthBench Professional and NOHARM, with the research preview Darwin at 72.8% on MedXpertQA, 82.7% on HealthBench Professional and 87.2% on NOHARM and a perfect score on MedQA. SourceOpenEvidence, businesswire.com/news/home/20260903878517/enread 2026-10-06 |
|
| Browser & Computer Use | Not documented |
|
Answering, drafting, coding and communicating all happen inside OpenEvidence's own app. OpenEvidence names no control of a browser, desktop or other screen. SourceOpenEvidence, businesswire.com/news/home/20260903878517/enread 2026-10-06 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Product updates
OpenEvidence introduced a new family of medical AI models, including Osler, Sackett, and Snow, which offer varying levels of research depth and response speeds for clinicians. The company also integrated a precision oncology knowledge base into its specialized oncology AI sub-agent to provide genomic interpretations.
Bears on: Agent capability
View sourcePricing
Free for verified clinicians
What is public
OpenEvidence's own releases state the production models are free for verified clinicians and that Coding Intelligence and the Dialer are available to all verified clinicians in the app. No enterprise or health system pricing is published; the July advertising CPM figures came from third party research and are removed.
Variable cost rationale
Individual clinicians pay nothing; enterprise or health system terms are not published.
Sales call required
Mixed (some tiers require a call)
Free / trial
Free for verified clinicians
Related vendors
- Abridge — AI platform for health systems that turns clinical conversations…
- Adonis — AI orchestration platform for revenue cycle management whose…
- AKASA — Generative AI for the hospital mid revenue cycle: Coding and CDI…
- Alaffia Health — Agentic AI for health plan claims operations: utilization…
- AlethianAI — Connected AI agents for medical practices share one patient record
- Ambience Healthcare — AI platform for health systems pairing an ambient scribe with…
Alternatives to OpenEvidence
The closest documented capability profiles to OpenEvidence among healthcare agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Fathom5.0 / 14Adds documented Security, Identity & Governance
- Glass Health5.0 / 14Adds documented Security, Identity & Governance and Memory & State Persistence, among others
- Marisa.Care5.0 / 14Fuller documented coverage on Integrations & Tool Calling and Workflow Orchestration
- Phluence5.0 / 14Adds documented Security, Identity & Governance
- Qventus5.0 / 14Adds documented Observability & Auditability
- Sunoh.ai5.0 / 14Adds documented Observability & Auditability and Browser & Computer Use
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded