Clinical Reference & Evidence
D

Dr.Oracle

Dr.Oracle is a medical question and answer assistant for physicians, residents, physician assistants, nurse practitioners and medical students. A user types a question in plain language and gets a narrative explanation with references to specialty guidelines, peer reviewed studies, FDA drug labels and case reports, each citation tagged with the strength of its evidence. Practicing physicians write and keep updating an internal library of guideline articles, and a Medical Advisory Board reviews generated answers on a regular cycle. Clinicians use it to look up a guideline, build a differential from case findings or check the evidence behind a management step. Research Mode appraises the strengths and weaknesses of up to 25 papers at once, and students use the same tool for pre rounding and question bank review.

TheDeep, LLC operates Dr.Oracle and calls it 100% physician owned and free of pharmaceutical funding. The company counts more than 53,000 medical professionals using it, and the Android app has passed 100,000 downloads since the service launched in January 2024. It runs in a browser and in iPhone, iPad, Apple Vision and Android apps on one account, in several languages. New users get a seven day trial, and refunds are honored within 30 days. The US App Store lists a $24.99 monthly plan beside lower priced tier items from $5.99. The terms prohibit entering protected health information, and no business associate agreement is offered.

AI Health Index verifiedNovember 19, 2026
Compare Dr.Oracle with other vendors
Founded
—
Headquarters
—
Categories
clinical-reference-and-evidence, clinical-decision-support, workforce-and-training
Indexed Products
Dr.Oracle AI Medical Assistant, Research Mode
Buyer Segments
Physicians and advanced practice clinicians, Residents, Medical and physician assistant students
Assessment

Capability Axes

An AI Health Index grade measures what a buyer can verify from public sources on the date shown. It is not a rating of how good the product is. A vendor can build an excellent system and grade low on an axis because it publishes nothing an outsider can check. How grades read

AI Capability
AA on AI CentralityThe artificial intelligence is the product. Remove the model and there is nothing left to sell.
Vendor Published

The product a subscriber pays for is the generated answer. A question in plain language returns a synthesized explanation with references to guidelines, journal studies, FDA drug labels and case reports, and a follow up question continues the same thread. Research Mode applies the same engine to appraisal, sorting findings, strengths and weaknesses across as many as 25 papers.

The physician written guideline library gives the model a curated base to draw on, and the differential feature turns case findings into a list of possibilities. Take the model away and the guideline articles remain, but they serve as inputs to the assistant rather than a reference product sold on their own, and every paid tier sells access to the synthesis.

DD on Autonomy and Oversight ModelNo oversight structure is published. An absolute claim that the system does not err grades here too, because a buyer who believes it will not build the review step that would catch a failure.
Vendor Published

A generated answer goes straight to the user in the conversation, with citations to open, and nothing requires a clinician to approve it before it is read or acted on. Oversight sits after the fact: the Medical Advisory Board reviews generated answers on a regular cycle, and shared articles can show a review status. The terms put verification on the user before any consequential action, and the homepage carries the feature heading 'Double Check Before You Treat'.

The same page sells the opposite under the heading 'Never Miss a Diagnosis' and states in its FAQ that the product has never had a report of hallucinations, while the App Store listing claims the highest USMLE score in history. Those claims invite trust without review, and the product supplies no review step of its own.

CC on Model and Technology TransparencyThe architecture is described in general terms with nothing identified. Proprietary is asserted rather than explained.
Vendor Published

The assistant takes a question in plain language and produces a cited response grounded in medical guidelines and published research, rating the evidence behind each citation. The company separates its physician written core guideline material from the answers the service generates, and Research Mode groups papers for structured appraisal. App updates describe continuing human review with weekly fine tuning, and version 2.0.3 brought an updated FDA drug database and multilingual support.

Marketing compares the product's exam scores with OpenEvidence, ChatGPT 5, Claude 4, Google Gemini and MediSearch, which names the competition rather than the engine. The model family behind the answers, what the weekly fine tuning trains on, and which app or web release corresponds to a model update are not identified.

CC on Model Supply Chain DisclosureThe architecture is described and no model provider is named. Naming a hosting provider alone does not lift a record out of this band. Record the host in the note, because it matters for residency and breach scope, and grade on the model layer, which is the question this axis is named for.
Vendor Published

Medical questions may go to external AI model providers to generate a response, and account information is stored on Amazon Web Services in the United States, as the privacy policy sets out. Google Analytics, Cloudflare, Sentry and Plausible handle analytics, traffic and monitoring, and Apple or Google sign in can supply identity. The App Store privacy label adds that usage data may be used to track users across other companies' apps and websites.

That covers much of the operational chain and the purpose of the model transfer. The model providers themselves go unnamed, and so do the terms that govern a question once it reaches them: how long it is kept and whether it can train the provider's models.

CC on Clinical and Operational EvidenceNamed customers, or vendor reported percentages with no method, denominator or reference standard. Scale of use is recorded here and is not treated as evidence of benefit.
Vendor Published

Adoption is the strongest figure: more than 53,000 medical professionals by the company's count, and over 100,000 Android downloads. Testimonials come from roles rather than institutions, among them an internal medicine resident, an obstetrician and a medical student who credits the app for clerkship honors. The headline performance claim is exam based.

The homepage shows scores of 100% on USMLE Step 1 and Step 2 and 97.5% on Step 3, citing Kung et al. (2023) for the exam method, and the company has shared a 2024 answer sheet for a sample licensing exam. Those tests assessed an earlier system on exam questions. No study ties the current release to answer accuracy in clinical use, time saved or patient outcomes, and no hospital or program is named as a deployment.

BB on AI Safety and PHI StewardshipCategorical commitments are published, such as no training on customer data, without the retention schedule behind them.
Vendor Published

The terms prohibit protected health information in questions, and an automated deidentification filter strips direct identifiers entered by mistake; the privacy policy says plainly that the filter can fail. A user can unlink question history from the account, though an unlinked question can still be reviewed or adapted for educational use. Account deletion removes personal information from company servers within 30 days.

Submitted content can be used for service improvement and quality review, which sits beside the weekly fine tuning described in app updates, and no limit is set on training with user questions or on retention at the model providers. There is no separate process for handling a clinically unsafe answer once one is reported.

Regulatory and Compliance
BB on HIPAA and BAA PostureBusiness associate status is stated and supported by a substantive privacy document, with the agreement or its scope not fully published. For a vendor outside the United States, an equivalent regime documented to this depth grades here. A vendor whose product does not process protected health information grades here too when it says so plainly and explains the scope, with a service privacy document behind it: stating a position a buyer can rely on is the posture this axis grades, and the band above is closed to it because there is no agreement to publish.
Vendor Published

The position is exclusion. The terms and the service privacy policy keep protected health information out of the product, and the operator states that it does not offer a business associate agreement. Questions are meant to be general medical and scientific ones, a case described without identifiers, and an automated filter removes identifiers submitted by accident, with the privacy terms warning that it is imperfect.

An anonymous identifier still links question history to an account until the user unlinks it, and questions can stay available for internal review. That is a clear stated scope backed by a service privacy document. It also leaves no contracting route, so a hospital or residency program cannot bring the tool inside its own HIPAA program.

DD on Security Certifications and Trust CenterControls are asserted with nothing independent behind them, or nothing is published. Read the note before concluding anything: this is the grade most often corrected on a second pass, because assurance material frequently sits on a parent domain or inside an old announcement rather than on the product pages.
Vendor Published

Account and question data sit on Amazon Web Services in the United States. The privacy policy commits to reasonable security measures, lists Cloudflare, Sentry and other monitoring services, and lets users unlink question history and request access, correction or deletion of their personal information. Apple and Google sign in are optional routes into an account. Those are the service's own operational controls, asserted without an outside check.

No SOC 2 report, ISO/IEC 27001 certificate, penetration test or trust center covers TheDeep's operation of Dr.Oracle, and Amazon's certifications belong to the host rather than to the service. There is no institutional plan with security terms of its own, so a site license would rest on the consumer privacy policy.

CC on FDA and Regulatory StatusNo device claim is made and the product is scoped accordingly. Most administrative and operational products sit here and are not penalized for it, because this axis grades the appropriateness of the positioning rather than possession of a clearance. A stated limit on use, such as research use only or a clause that the product gives no medical advice, beside marketing that sells the product for clinical decisions also sits here, with both quoted, because the buyer has to ask which one governs.
Vendor Published

The terms define the service as an educational and informational tool for general medical and scientific questions and disclaim use as a substitute for medical advice, diagnosis or treatment, and the App Store listing calls it a medical education tool rather than a physician substitute. The product also sells clinical uses past that line.

One feature takes case findings and returns a differential, the app listing describes guideline based management options, and the homepage invites clinicians to check the evidence before treating. Both positions stand side by side on the company's own pages. No FDA clearance, exemption or regulatory pathway is named for the differential feature, and verification and every decision are left to the user.

BB on AI Governance and Bias DisclosureA governance framework with named process behind it, such as certification to an artificial intelligence management standard, or material written for a customer own review committee to evaluate the product with. A framework covering the vendor whole portfolio counts when its scope reaches this product; what it cannot supply is the product specific evaluation the band above asks for. Published results within the one group where the performance of this product is most at risk also sit here: evaluation on the question this axis asks, short of the comparison across groups the band above requires. A medical device clearance or certification, and the quality system behind it, does not reach this band on its own; it is graded under regulatory status. A regulator reviewed plan governing how the model may change does reach it.
Vendor Published

Practicing physicians write and maintain the core guideline articles, a multidisciplinary Medical Advisory Board reviews generated answers, and the public article workflow marks material awaiting review apart from approved material. The editorial process puts review around changes to reference content and around a sample of generated answers, and the company states that it takes no pharmaceutical funding and is 100% physician owned. That is governance of sources and publication.

It does not reach model behavior: despite claims of unbiased evidence and coverage of every medical field, there are no results by specialty, by patient demographic or by type of question, and none for the languages added in version 2.0.3. The advisory board's members are not named.

CC on AI Liability and RecourseMechanisms exist that let someone challenge an output, such as audit trails, source traceability or review before commit, with nothing standing behind the output and no route for the harmed party.
Vendor Published

Inline references let a user compare a generated explanation with the guideline or study it cites, which is the one check on an answer. The agreement assigns interpretation and verification to the user, disclaims any warranty that errors will be corrected, excludes indirect and consequential damages, and requires the account holder to indemnify TheDeep for specified claims.

The money back route is commercial: refunds are honored within 30 days, and the refund policy governs how a trial converts. None of that attaches to an error in a differential or a medical explanation, so a wrong answer carries no correction duty or remedy, even though the marketing promises a product that never misses a diagnosis.

Integration and Deployment
DD on EHR and Interoperability DepthNo integration evidence. A connector described as available on request grades here until one exists.
Vendor Published

Access runs through a web browser and the iPhone, iPad, Apple Vision and Android apps, with one login across devices and optional Apple or Google sign in. A clinical question enters through the chat window, typed by the user, and the answer and references come back in the same conversation; Research Mode presents its appraisal there too. Nothing moves between the assistant and a clinical system.

There is no EHR connection for importing chart data, no route for writing a finding back to a record, no launch from inside an EHR and no institutional deployment in a named system. Clinicians who use it at the point of care do so in a separate app beside the record.

BB on Deployment Model and Data ResidencyOptions and residency are stated with isolation or the processing path left open.
Vendor Published

The service is hosted only, on Amazon Web Services in the United States. Users in other countries reach it through a browser or the apps, account information is transferred to and processed in the US, and the privacy policy covers safeguards for international transfers where the law requires them. The app sells worldwide, with listings in Chinese, Arabic, Spanish and Russian storefronts, so many users sit outside the country where their data rests.

Questions may also go to outside model providers whose locations are not given. There is no private hosting, regional option or institutional tenant, and nothing describes separation between one subscriber's history and another's beyond the account itself.

Commercial
BB on Commercial TransparencyA price or a pricing basis is published without full tiers, so a buyer can size the cost before making contact.
Vendor Published

Dr.Oracle sells individual subscriptions after a seven day trial, with refunds honored within 30 days. The US App Store lists a $24.99 monthly subscription, $63.00 for three months and $239.00 for a year, and its purchase list also carries lower priced items for named tiers, from $5.99 and $9.99 up to $12.99 a month and $124.99 a year, plus a monthly plan at 30% off for $17.49, without saying what each tier includes.

Web subscriptions are managed on the account's billing page, and the refund policy says a trial bills at the plan selected at signup. There is no published web price list and no institutional or residency rate, so the self service route is clear while the tier structure is not.

CC on Setting and Specialty CoverageCoverage is claimed broadly without specifics, or stated clearly with nothing validating it yet.
Vendor Published

The service is built for medical students, residents, physician assistants, nurse practitioners and physicians, and its case discussion and literature features cover internal medicine, pediatrics, emergency medicine, obstetrics and gynecology and oncology among other fields. It serves two settings: point of care reference during clinical work, and study, from pre rounding to question bank review.

The testimonials match that split, with a hospitalist and an obstetrician on the clinical side and a resident, a medical student and a physician assistant student on the education side. Adoption is described across hospitals, residency programs and universities without naming any. Nothing validates the current release in inpatient rounds, outpatient consultation or a training program.

Alternatives

Alternatives to Dr.Oracle

Other products that do the same kind of work as Dr.Oracle, and what each one does.

Clinical search and answer platform for UK clinicians and students, from Medwise AI Ltd, registered in England and Wales and based in Leeds, having started in Cambridge.

NeuroTrials.ai is a curated database of practice changing neurology trials, built by board certified neurologists at NeuroCare Medical LLC in Florida, with an AI assistant that answers clinical questions from it.

Clinical AI reference platform from CaryHealth (Cary Rx Incorporated, Washington, DC), launched in August 2024 on the back of the company's GalenAI acquisition and available at askclair.ai, on the web and as iOS and Android apps, with an API for embedding.

Doximity Ask, formerly Doximity GPT, is the clinical AI assistant inside Doximity's professional network, whose members include more than 85 percent of US doctors.

AMBOSS AI Mode Clinical Care answers medical questions with short summaries linked to selected sources. It draws on the edited AMBOSS library, US clinical guidelines and a drug database, adapting the presentation to a clinician's specialty and answering in the language of the question.

Generative chat interface over the American Society of Clinical Oncology's own library of clinical practice guidelines, announced in May 2025 and developed by ASCO and Google Cloud on the Vertex AI platform with Gemini models.

Commercial

Pricing

Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.

Entry Price Pricing Basis BAA Tier Implementation Source
$24.99/month (US App Store)
$24.99 baseline
Individual subscription after a seven day trial. The US App Store lists $24.99 monthly, $63.00 for three months and $239.00 yearly, plus lower priced tier items from $5.99 to $124.99 with no feature comparison. The terms exclude protected health information and say a business associate agreement is unavailable. Download or web signup is self service. No separate setup charge applies to the individual subscription. Vendor Published

The trial converts to the plan selected at signup unless cancelled, and refunds are honored within 30 days. Web subscriptions are managed on the account's billing page and App Store subscriptions through Apple. No institutional or residency rate is listed.