Beta · Launching Spring 2027
ORPHEON
Clinical Intelligence Platform

Ask the question.Run the study.Advance medicine.

The first of its kind clinical intelligence platform, built by physicians, for clinical practice and research. Hypothesize. Interpret. Innovate. That's what clinicians and researchers do best. Orpheon, powered by Cantos, handles everything in between.

One question, from the literature to a manuscript
Illustrative example · demonstration data
Discover
Step 1
anastomotic force and continence after RALP
47 papers · sorted by relevance
Urology
Factors Associated With Recovery of Urinary Continence: A Multicenter Comparison of Pelvic Fascia-sparing and Standard Robotic-assisted Radical Prostatectomy
Wald et al. · Eur Urol Focus · 2025 · n=1,155
Urology
Preservation of the Neurovascular Bundles Is Associated With Improved Time to Continence After Radical Prostatectomy but Not Long-term Continence Rates: Results of a Systematic Review and Meta-analysis
Reeves et al. · Eur Urol · 2015 · systematic review
General surgery
A Pilot Study on the Impact of Da Vinci 5's Force Feedback on Clinical Outcomes: Does It Lead to Less Tissue Trauma in Colorectal Surgery?
Chang et al. · Am J Surg · 2025 · n=68
Cantos · Research opportunity
Force is now being measured in surgery. It has not been measured against time to continence.
Vault
Step 2
Prostate Cancer
1,247 patients
One of 7 databases · 5,359 patients in this Vault
Continence recovery
49.4%
3 months
68.6%
6 months
98.0%
12 months
Published at 12 months: 84 to 97% · Ficarra 2012
Force at the anastomosis
Newtons · n=571
Mean force2.2 N IQR 1.7 to 2.5
Peak force4.3 N IQR 3.3 to 5.1
No published benchmark yet
Cantos noticed · Pattern
A relationship between force applied at the urethrovesical anastomosis and time to continence.
n=296 in subgroup · q<0.001 · replicated on held-out halves · found without being asked
Build a study
723 tests run · 51 credible findings · 7 replicated
Studio
Step 3
Plan Ground Write RefineSubmit
Anastomotic force and time to continence after robotic prostatectomy.
Figure 1 · Time to continence
% continent
050100036912Months after surgery
PrespecifiedKaplan-MeierCox model
Two force groups · 571 of 1,247 patients · median follow-up 41.8 months · drawn from your Vault
Reviewed before you submit by a methodologist, a statistician and a clinical reviewer.
Journal fit · 15 journals read
Paper strength 76 / 100, weighed against each journal
The Journal of UrologyMatch72
European UrologyReach58
Find the gap.
Cantos reads the literature and shows what has not been answered.
Find the signal.
Cantos tests your cases for patterns no one thought to look for.
Write the paper.
Draft from your data, review with Cantos, and match to a journal.

Cantos is the intelligence layer behind all three steps. Specialty-trained and shaped by your peers.

Meet Cantos
The intelligence layer

Meet Cantos.

Clinical intelligence, curated for your specialty.

We're building Cantos one specialty at a time. Each discipline starts with an independent foundation: the specialists who practice it, the literature that defines it, the outcomes that measure it. Every answer Cantos returns is shaped by your field's evidence, not averaged across all fields.

40M+
Papers indexed · Updated continuously
PubMed · ClinicalTrials.gov

Trained by specialists

Practicing specialists are shaping Cantos's reasoning one discipline at a time: how the field thinks, which evidence matters, how its journals evaluate work.

Grounded in specialty evidence

Every peer-reviewed paper indexed in PubMed, with your specialty's registered trials from ClinicalTrials.gov. Every clinical answer is anchored in your specialty's own evidence and cited back to the source, so you read the original, not a paraphrase.

Speaks your field's language

Cantos answers in your specialty's voice and writes to your journals' conventions, not the hedged prose of a general-purpose chatbot.

Why Orpheon

Innovation in every layer.Engineered to transform medicine.

The path from observation to publication is broken. Not for lack of insight, not for lack of data, but because every layer of the research workflow was built for a different era.

01 / 06

First of its kind. Built for clinical practice and research.

Built for a different era

Existing clinical AI tools search the published literature to answer clinical queries.

With Orpheon

Orpheon is not a search engine. It is the first platform designed to measure your own outcomes against that evidence, identify the gaps and the questions yet to be answered, and carry the work through to publication, transforming clinicians into medical innovators.

Built on Cantos

Three products.One methodology.

The gap you find in Discover becomes the question you answer in Vault, where Cantos cross‑references your data against the same literature to flag when it's mature. The analysis you run in Vault becomes the manuscript you draft in Studio, with citations pulled from Discover. One continuous session, three surfaces, one body of work.

01
Discover.

Answer any question. Find every gap. Measure your practice.

Ask your question in plain language. Cantos searches the published literature, ranks every study by relevance, synthesizes the evidence, and turns the gaps into your next research opportunity.

PubMed and ClinicalTrials.gov
Searched in one pass
Relevance Score
Every study ranked
Cantos synthesis
Every claim linked to its source
Literature Monitor
Keeps watching after you ask
orpheon / discoverIllustrative
1Ask2Answer3Your practice
ClinicalResearchYour Vault is linked
Is single port partial nephrectomy associated with improved clinical outcomes?
Your practice against the evidence
Warm ischemia time
24min
Cited 14 to 22
Length of stay
1.2d
Cited 1 to 2.5
Complications
4.4%
Cited 2 to 9
Positive margins
2.9%
Cited 1 to 6
Bottom line
Nguyen 2024Hu 2024+14 more
The evidence
Length of stay
1 to 2.5 days · 9 cited studies
Your practice
1.2 days
Single port cases · n=54
Cantos lines up each measure the cited studies report with the same measure from your Vault, and keeps the row current as cases are added.
Cantos noticed a divergence
Your single port cases and the cited studies differ on warm ischemia time. Open the comparison to see where.
  • Natural language search. Ask the question the way you'd ask a colleague. Cantos searches PubMed and ClinicalTrials.gov in one pass.
  • Relevance Score. Every study is ranked by how directly it answers your question, weighted by evidence quality, study design, and recency. The highest-yield reads sit at the top of the list.
  • Cantos synthesis. The top studies are read, summarized, and coalesced into an evidence-based answer written in the voice of your specialty, with every claim linked back to the source paper.
  • Research opportunity generator. Cantos turns previously unidentified gaps in the literature into your next high-impact research opportunity, with novelty, impact, and feasibility scored. One click carries it into Studio.
  • Cohort feasibility, instantly. Cantos checks your Vault against every surfaced research opportunity, telling you how many matching cases you already have and how soon you'll hit a publishable target N at your current case volume.
  • Literature Monitor. Cantos keeps watching the field after you ask. A new trial against your question, a competing series moving the standard, a gap that begins to fill: Discover flags it on your dashboard the morning it hits.
orpheon / vault / prostate-cancerIllustrative
1A case arrives2The monitor updates3The curve draws
Primary database
Prostate Cancer
642 casesLive
Case IDTreatmentTx datePathologyRecent PSA
OP-2231
RALP
10/02/26
3+4=7
<0.1
Summary
Pad-free at 3 months, PSA <0.1.
OP-2214RALP09/18/264+3=7<0.1
OP-2197RALP09/04/263+3=60.12
OP-2183RALP08/21/263+4=7<0.1
Metric monitor
3-month continence
52.1%n=612
+0.2 pp
12-month continence
94.6%n=570
No change
Positive margin
11.8%n=588
No change
30-day readmission
3.1%n=641
No change
Continence recovery
Time to event
100%004812 mo
Published benchmark
RALP
Continence at 12 months
94.6% your VaultPublished 84 to 97%
Ficarra 2012 · primary source, reviewed
Patient names and MRNs are encrypted on this device and are never sent to Cantos.
02
Vault.

HIPAA-aligned. Clinician controlled. Built for practice and research.

One HIPAA-aligned database for every study you run: retrospective series, prospective clinical trials, QI projects, and long-term outcomes tracking. Capture cases by voice, photo, or text input, or import an existing spreadsheet in bulk. Cantos runs your statistics on demand and reports them in plain language. You control the methodology. You own the data.

HIPAA-aligned
BAA available on request
PHI on-device
Encrypted in transit and at rest
Never sold or shared
PHI never sent to third-party AI
Research is opt-in
Revocable, never includes PHI
  • No schema to design. Pick your specialty and your procedures, and the right fields are already there. Vault ships with your curated specialty dictionaries, so you never build a data dictionary or guess your columns up front.
  • One living database, not scattered projects. Enter each patient once. Your studies are saved cohorts over that single corpus, and they keep growing as you log cases. De-identified by design.
  • Capture in seconds, at the bedside. Log a case by voice, by photo of an op note or lab report, or by typing. Cantos structures every entry against your fields. Patient-reported outcomes are fully managed by Vault: it schedules each survey, sends it, and files the response back to the right case, with no added work for you.
  • Cantos-guided statistics, clinician directed. You set the methodology. Cantos runs the analysis, from descriptive summaries to group comparisons, adjusted multivariable models, and survival curves, and translates every result into plain language. The biostatistics you would normally wait on a statistician for, ready the moment your cohort is.
  • Knows when your data is ready. Vault tracks your accumulating series against the literature you queried in Discover, tells you how close you are to your target N, and flags the moment a finding becomes publishable, then hands it to Studio.
  • See where you stand. Compare any outcome two ways: against your own historical baseline and against published literature, before a reviewer asks.
03
Studio.

From protocol to manuscript. Built to your specifications. Integrated across Orpheon.

Studio is Orpheon's output engine, where your clinical data and research become a polished product. Your Vault and Discover libraries feed the content. You interpret the data. Cantos acts as your research assistant, drafting your IRB protocols, research abstracts, and manuscripts. The work moves at your pace, in your voice, under your authority.

IRB protocols
Designed from a Discover opportunity
Abstracts and manuscripts
Drafted from your Vault data
Critical reviewer
Reads the manuscript end to end
Journal fit
Ranked as the manuscript improves
orpheon / studio / manuscriptIllustrative
1The path2The outline3The draft
Manuscript · retrospective series
Stage 7 of 10 · Draft
Same-day discharge after robotic partial nephrectomy: a retrospective series
Outline · Introduction
All reviewed
Disease burdenOpen with how often small renal masses are treated and how often partial nephrectomy is performed.[1][2]
Evidence gapState that same-day discharge has not been reported with standardized follow-up.[6]
Draft · Results ¶1
Every number bound to your Vault

The registry held 412 patients before exclusions, of whom 38 did not undergo the index procedure, leaving 374 patients for analysis. Median follow-up was 26.4 months , and 341 of 374 patients ( 91.2% ) were followed at least 12 months.

Revisions on this paragraph · 1
Report follow-up completeness in the first paragraph
Done
  • Every output, one engine. IRB protocols, research abstracts, and manuscripts all draft from the same Vault data and Discover library, in the same continuous workflow.
  • Publication-grade statistics, built in. Studio runs the analyses your manuscript actually needs (Kaplan-Meier survival, Cox regression, subgroup analysis, multivariable models) and renders the tables and figures alongside the text that interprets them. No third-party software, no switching tools.
  • Cantos learns your voice. The more you write, the more Cantos learns: your phrasing, your structure, your way of interpreting a result. Cantos is the research assistant; you are the principal investigator on every page. The integrity of the work remains yours.
  • Protocol to manuscript, one continuous workflow. Studio designs your protocol from a Discover opportunity. As Vault accumulates the data, Cantos drafts your manuscript outline and content section by section, guided by your input and anchored with citations from your Discover library. You approve every milestone.
  • Critical reviewer and journal fit, on demand. Cantos reads your manuscript end to end for alignment and identifies the journals most suited for submission based on content, methodology, novelty, and previous publication trends. Run the critical reviewer pass to tighten the work, then watch the journal ranking recalculate as your manuscript improves.
  • Submission-ready, journal-formatted. Pick the target and Studio formats your manuscript to spec: reference style, structured abstract, figure captions, cover letter, and supplementary files. Studio takes the busy work out of the submission process.
Dashboard

Your command center.

Cantos reads across Discover, Vault and Studio and puts the next step first: the manuscript nearest to submission, the cohort that reached its target, the new papers in your field.

Orpheon CommandCantosLive
Good afternoon,
Doctor.

Your manuscript is one stage from submission. A cohort is ready to investigate.

Ask CantosAnything across Discover, Vault and Studio
Cantos recommendsNearest to ship
90%
Your manuscript is one stage from submission.
In Revise · illustrative manuscript
Also ready
A cohort reached its target, 512 of 500Investigate
64 new papers across 2 monitorsReview
Three cohorts have no target yetSet target

Research map.

6 cohorts · 2 bridges

How your work connects: cohorts, the concepts they share, and the studies that point to them.

Prostate cancern = 1,180
Kidney cancern = 960
Bladder cancern = 410
BPHn = 512 / 500
Urolithiasisn = 1,520
Ureteral stricturen = 160
Perioperative blood loss3 cohorts
Length of stay2 cohorts
Vault cohortConceptCross-cohort bridgeDiscover
6
Cohorts
10
Saved studies
1
Manuscript
8
Stored links
Illustrative example · demonstration data
From the founder
Christopher Russell, MD, FACS · Founder, Orpheon Health

Too many good ideas die in case logs.Most never reach the literature.

Every clinician I know has research that never makes it out of their case logs. I've spent my career as a clinician and a researcher, and I've watched it happen to colleagues I respect and in cases I've cared about. Insight worth publishing, observations worth studying, lost to records nobody else will ever read.

Literature is fragmented. Outcome data is locked inside records built for billing. Statistical software was made for statisticians, not clinicians. Manuscript drafting takes months. And the toolkit to make any of this efficient and accessible for practicing clinicians has never been built.

Insight worth publishing, observations worth studying, lost to records nobody else will ever read.

Orpheon is being built to close that gap. In Discover, you see what the literature already knows, and where it falls silent. In Vault, you capture your data prospectively, structured for research from the first case, with statistical analysis designed for clinicians, not statisticians. In Studio, you seamlessly draft the protocol, abstract, and manuscript. Cantos, the clinical intelligence model behind all three, is trained one specialty at a time on the literature that defines it.

This is the platform I wished I had as a resident, the platform I needed as a young attending, and the platform I am building now as an experienced surgeon. For every clinician who has ever seen something in clinic worth studying, I'd be honored to have you with us as a founding member.

Christopher Russell, MD, FACS
Founder, Orpheon Health
Become a founding member

Join the beta.

We're hand-selecting the first 100 clinicians and researchers to help shape Orpheon before public launch. Discounted founding member pricing during beta. No commitment. Have a specialty you'd like Orpheon to support? Tell us below.

How would you like to engage?
Which products are you most interested in? (check all that apply)

We'll only email you about Orpheon. No spam, we never sell your data. Unsubscribe anytime.
By joining, you agree to our Terms and Privacy Policy.

FAQ

Questions & answers.

When will the beta be available?

Beta access opens spring 2027. We're rolling out specialty-specific waves beginning with the surgical subspecialties, with additional specialties added in subsequent waves. Sign up now to reserve your spot.

Who can join the beta?

Any practicing clinician, researcher, or trainee working in a surgical specialty, including attendings, fellows, residents, medical students, advanced practice clinicians, and PhD researchers. Additional specialties will be added in subsequent waves. You don't have to be actively publishing today. If you've ever seen something in clinic worth studying, you're exactly who we're building for.

What if I don't consider myself a researcher?

Most of the clinicians we're building this for don't, at least not yet. Orpheon is built for every clinician, regardless of practice setting or research experience. For the active clinical researcher, Orpheon streamlines your workflow and amplifies the impact of your work. For the curious clinician, Orpheon is the gateway to translating your ideas, observations, and personal experience into measurable impact. Whether you want to compare your outcomes with the published literature in your specialty, draft an abstract, write a case report, or publish a manuscript, Orpheon generates the information and produces the outputs you need. Orpheon grows with you; there's no day-one expectation of being a researcher.

Is it free?

Alpha testing is free and by invitation. Beta is paid, at a discounted founding member rate.

Is there a mobile app?

Orpheon's web app runs on any modern browser, including mobile.

Is my patient data safe?

Vault is engineered as a de-identified data layer. Patient identifiers are stripped on-device before anything reaches Orpheon's servers, so no PHI ever leaves your device. You control what goes in, and you retain ownership of the data you contribute.

Can I export my data?

Yes, always. Your data is yours. Vault registries export in standard formats (CSV, JSON), and Studio outputs (manuscripts, abstracts) export in DOCX or PDF. There's no lock-in: if Orpheon isn't for you, you can take everything you've built.

How is this different from other clinical AI tools?

Orpheon is not a search engine. Other clinical AI tools summarize what's already published. Orpheon is a first of its kind platform for clinical practice and research: Cantos, Orpheon's proprietary clinical intelligence model, answers your question, measures your own outcomes against the evidence, identifies what's missing in the literature, and assists you in drafting the manuscript. Literature, data, manuscript, end to end.

One platform that compounds with every case.

Beta members receive discounted founding member pricing.

Join the beta