Secure upload
Upload one file or a full document set from the workspace.
Biomedical document intelligence platform
Turn Biomedical Documents IntoVerifiable Intelligence.
Store documents. Find the facts. Follow the patient story visually. Build deterministic reasoning. Ask questions. Get answers with actual evidence. Release sensitive documents safely for research.
- Notes, reports, PDF, FHIR, HL7 and more
- Facts keep status, time and source
- 141 documented REST endpoints
- Local AI by default
Interactive implementation example
clinical-note.txt · de-identified preview
On this page
01
Prepare
Secured Document Storage
Bring every document in. Keep every original protected.
Upload a file, send a large batch, or connect an existing ETL pipeline through the API. BioMedora stores originals with encryption, organization isolation, role-based access, and a complete audit history.
Bring documents in your way
API and SDK
Send documents from products, EHRs, LIMS, and internal applications.
Existing ETL pipelines
Connect scheduled jobs, queues, webhooks, and data-lake workflows.
Originals stay protected
Immutable source, retention policy, checksum, and complete audit history
PDFDOCXTXTCSVEmailFHIRHL7 v2C-CDA
Role-based access
People see only the documents and services their role allows.
Organization isolation
Every document, result, and action stays inside its organization.
Controlled originals
Download, export, and release rights are separately governed.
02
Understand
Fact Extraction
Read each document once. Use it everywhere.
BioMedora extracts structured facts, dates, relationships, and source evidence. Then it turns them into services every authorized user and application can reuse.
Workspace Intelligence
Five working views. Every one tied to the source.
Move from structured patient data to timelines, connected evidence, search, and cited answers without leaving the workspace.
01
Patient Data
Facts keep their meaning: present or absent, current or past, coded, dated, and tied to the source.
metformin · current · RxNorm 6809 · p.4
02
Patient Timeline
Events from many documents become one cited patient story. Negated and family-history mentions stay separate.
2024-03-04 · medication started · [DOC-18472]
03
Care Graph
Conditions, drugs, labs, variants, procedures, and documents connect without losing their evidence.
Medication → TREATS → Condition · source attached
04
Evidence Search
Search by words, biomedical concepts, relationships, or time—and retrieve the exact supporting passages.
query → ranked passages → evidence pack
05
Evidence Chat
Every answer is checked claim by claim. Numbers, citations, support, and uncertainty stay visible.
SUPPORTED · 96% · citations [1][2]
Deterministic first
Three safeguards. One dependable result.
Prepare the document. Capture the facts. Check the AI. Each step has one clear job.
01 · Document intelligence
Prepare every format
Classify files, find meaningful sections, read tables, preserve page positions, and route scans through OCR.
PDFDOCXFHIRHL7 v2C-CDACSVEmailText
02 · Clinical meaning
Capture what was actually said
Find facts, negation, who experienced them, when they happened, and how they relate. Then connect standard biomedical codes.
SNOMED CTRxNormLOINCICD-10-CMMedDRAHPOHGVSATCUCUM
03 · AI validation
Check AI before delivery
Choose deterministic, local, private, or approved external AI. Verify evidence, numbers, and safety before a claim reaches the user.
Local AI defaultClaim verifierSafety validatorUncertaintyAudit
Pipeline Services
Follow one document through all 12 stages.
Open any stage to see what happens, why it matters, and which API runs it.
Clinical NLP
- What it does
- A deterministic engine — medspaCy TargetMatcher + ConText with an in-house lexicon fallback — extracts medications, conditions, labs, procedures, genes and more, with dosages, routes, reference flags, and HGVS variants.
- Why it matters
- Repeatability you can audit. The same document always yields the same facts — no sampling temperature involved in establishing what a document says.
- Example input
- Example output
Started metformin 500 mg BID. Creatinine 1.4 mg/dL (high)
{ "text": "metformin", "type": "Medication",
"attributes": {"dose":"500 mg","frequency":"BID"} }Exposed via the API playground — try it live.
03
Connect
Visual Analytics
Turn extracted facts into a story people can see.
BioMedora links events across time and connects patients, conditions, drugs, tests, and source evidence. Teams can explore the same intelligence without reading every document again.
Patient Timeline
See the patient story—not a stack of charts.
Years of events appear in order. Open any event to see its document, page, section, date, and source sentence.
Patient in focus
DEM-001 · 6 source documents
Date range
2019–2024
Event types
6
Cited sources
6
Filter
2019
2020
2021
2022
2023
2024
Accessible event list (screen-reader alternative)
- — Diagnosis: Type 2 diabetes diagnosed. HbA1c 8.1% at presentation; lifestyle modification advised.
- — Medication: Metformin started. Metformin 500 mg PO BID; tolerated without GI upset at follow-up.
- — Lab: Creatinine elevated. Creatinine 1.6 mg/dL (reference 0.6–1.2); repeat recommended in 90 days.
- — Procedure: Screening colonoscopy. No polyps; next screening interval 10 years.
- — Imaging: Chest X-ray finding. Mild cardiomegaly noted; comparison with prior unremarkable study.
- — Follow-up: Endocrinology follow-up. HbA1c improved to 6.7%; metformin continued; annual diabetic eye exam scheduled.
Care Graph
See what changed—and what connects.
Explore patients, medications, conditions, labs, and evidence. Every connection keeps its date, confidence, and source.
Accessible structured representation
- Patient —[HAS_CONDITION]→ Type 2 Diabetes
- Patient —[TAKES_MEDICATION]→ Metformin
- Metformin —[indicates]→ Type 2 Diabetes
- Patient —[HAS_RESULT]→ Creatinine 1.6
- Patient —[HAS_FINDING]→ Cardiomegaly
- Ibuprofen —[INTERACTION_RISK]→ Metformin
04
Reason
Reasoning with Evidence
Build deterministic reasoning. Get answers you can defend.
BioMedora applies rules first, checks every generated claim, verifies numbers, cites the actual evidence, and says what remains unknown.
Ask a question
Simulated interface mirroring the production contract of POST /v1/reasoning/query. Run it against your own tenant in the playground.
Two independent signals support a decline in renal function between 2020 and 2021, with an important caveat about follow-up:
Creatinine rose from within normal limits (1.1 mg/dL, 2020-02) to 1.6 mg/dL flagged HIGH on 2021-09-22.
The 2021 report explicitly recommended a repeat renal panel in 90 days.
- PARTIALLY SUPPORTED62% confidence
Metformin dosing was not re-evaluated in retrieved documents after the creatinine increase.
Absence of evidence in the current repository — not evidence of absence.
Uncertainty: No documents dated after 2021-12 were found referencing repeat renal testing — flagged as missing follow-up by data-quality analysis.
Evidence
[1] DOC-19011 · p.3 · Labs
“Basic metabolic panel: creatinine 1.1 mg/dL (normal).”
[2] DOC-20455 · p.3 · Labs
“Creatinine 1.6 mg/dL — HIGH. Repeat renal panel in 90 days.”
05
Protect
Privacy Studio
Automate large batches. Review every release that needs a human.
BioMedora finds PHI and PII, creates stable pseudonyms, and shifts dates by one patient-consistent offset. Reviewers can add, edit, or remove highlighted entities before approving a selected document for research release.
Privacy Studio
Research release batch · 12,486 documents
9,417 automated284 need review
DOC-04219 · Clinical follow-up note
Automated pass complete · Manual release review required
Original document
Protected source · read-only
() returned on after starting metformin 500 mg twice daily on . reports improved energy. Reviewed by . A1c decreased from 8.1% to 6.9%. Home address: . Continue therapy and repeat labs in 90 days.
Proposed release
Automatically de-identified · editable
Research Subject R-1042 (Study ID S-77-1042) returned on June 6, 2025 after starting metformin 500 mg twice daily on March 28, 2025. Research Subject R-1042 reports improved energy. Reviewed by Dr. Elaine Foster. A1c decreased from 8.1% to 6.9%. Home address: [ADDRESS REMOVED]. Continue therapy and repeat labs in 90 days.
Stable pseudonym
Every mention maps to R-1042
Dates stay meaningful
One +84-day patient offset
Residual PHI scan
1 possible entity needs review
Release status
Waiting for human approval
Medication, diagnosis, dose, sequence, and time intervals are preserved
06
Put to work
One Platform for Users & Developers
Give every organization one controlled way to use the intelligence.
People can work in BioMedora. Developers can connect through the API and SDK. Each organization decides who can see documents, use services, review releases, and manage the platform.
BioMedora Workspace
Move from document to decision in one place.
Work across Patient Data, Patient Timeline, Fact Lens, Care Graph, Evidence Chat, Evidence Search, Privacy Studio, and Usage & Storage in the functional frontend.
app.biomedora.com/demo
Live demoOverview
acme · Clinical reviewer
Search BioMedoraK
CL
clinician-02
Clinical reviewer
Clinical review workspace
Curate data to get evidence-backed insights.
Validate extracted facts, follow the patient story, and ask questions without losing the source context.
Evidence pipelineOperational
Ingest
Protect
Extract
Connect
Patients in scope
1,284
Tenant-isolated records
Source documents
18,204
Originals preserved
Privacy attention
12
Documents need review
Evidence questions
847
4,153 included requests left
Work queue
LIVEValidate extracted factsOpen Fact Lens
Review the patient timelineOpen Patient Timeline
Resolve privacy reviewsOpen Privacy Studio
Recent evidence
View alldischarge-summary-0412.pdfP-204 · today
endocrine-follow-up.pdfP-771 · yesterday
labs-export-q1.csvP-204 · Aug 26
Organization Support
One organization. Many roles. The right access for each.
Health systems, life-sciences teams, CROs, and product companies can bring their own users, policies, and workflows into one governed platform.
One document intelligence platform for the whole organization
Each organization controls its users, roles, documents, services, and policies.
Health systems
Patient timelines, evidence search, chart review
CliniciansCare coordinatorsData stewards
Life sciences
Cohorts, safety evidence, de-identified releases
ResearchersSafety analystsClinical operations
CROs and product teams
APIs, SDKs, batch workflows, governed workspaces
DevelopersProduct ownersSponsor teams
BioMedora platform
A secure home for documents, people, policies, and platform services.
Shared services, controlled by role
- Encrypted document storage
- Fact extraction and visual analytics
- Evidence-based reasoning
- De-identification review and release
- Workspace, API, SDK, usage, and audit
A researcher can review de-identified data without gaining access to original patient documents. A developer can use an API without receiving workspace administration rights.
Organizations and missions
Different organizations. One governed platform.
Health systems, life-sciences teams, research groups, and CROs give each user the tools their role needs—without creating separate data silos.
Healthcare
- Longitudinal patient intelligence across every chart, note, and feed
- Chart-prep and referral briefings generated with per-line citations
- Medication reconciliation with interaction and allergy conflict review
- Data-quality surveillance: contradictions, stale meds, missed follow-ups
Pharma
- Evidence extraction from labels, protocols, and regulatory documents
- Drug intelligence normalized to RxNorm / ATC with interaction knowledge
- Safety signal workflows over adverse events linked to source text
- Literature-grade reasoning with every claim verified against evidence
Biotech
- HGVS variant extraction with gene inference from unstructured reports
- Biomarker and phenotype intelligence via HPO-linked concept graphs
- Knowledge discovery over provenance-carrying research documents
- APIs that slot into existing LIMS and analysis pipelines
Clinical Research
- De-identified release of clinical documents for secondary research
- Cohort exploration over structured events instead of chart review
- Trial screening support with assertion-qualified eligibility facts
- Reproducible, auditable transformations for regulatory scrutiny
CRO
- Automated extraction across sponsor document formats at volume
- Structured deliverables mapped to your internal data schemas
- API-driven workflows with idempotency, batches, and signed webhooks
- Tenant isolation suitable for multi-sponsor operations
Your workflow here
If your organization moves biomedical documents, BioMedora can compile them. Tell us about your data and we will map the pipeline to it.
Talk to BioMedora
Developer SDK
Add BioMedora
to the tools you already use.
Connect an EHR, LIMS, research platform, data lake, or AI assistant through one API. Use 141 documented endpoints, clear error responses, live streaming, and signed webhooks.
- STEP 1
Create an API key
Choose the access scope. The secret is shown once and stored as a secure hash.
- STEP 2
Install the SDK
Use typed Python resources for documents, patients, privacy, reasoning, billing, and more.
- STEP 3
Send a document
Submit it once. Track every processing stage through the jobs API.
- STEP 4
Use the result
Receive facts, events, timelines, graph links, and source citations.
Python SDK · biomedora_sdk5-minute integration
pip install biomedora-sdk
from biomedora_sdk import Client
client = Client(api_key="YOUR_API_KEY",
base_url="https://api.biomedora.com")
# 1. Compile a document into structured intelligence
doc = client.documents.create(
content=open("discharge_summary.txt").read(),
patient_id="patient_123",
)
# 2. Read the cited longitudinal timeline
timeline = client.patients.timeline(patient_id="patient_123")
# 3. Ask an evidence-grounded question
answer = client.reasoning.query(
patient_id="patient_123",
question="What changed after the medication was started?",
)
for claim in answer["claims"]:
print(claim["status"], claim["confidence"], claim["text"])Also distributed as biomedora-client — identical API.
cURL
curl -X POST https://api.biomedora.com/v1/nlp/analyze \
-H "Authorization: Bearer $BIOMEDORA_TOKEN" \
-H "Content-Type: application/json" \
-H "X-Request-ID: req-0042" \
-d '{
"content": "Started metformin 500 mg BID. No evidence of pneumonia."
}'TypeScript — REST
// TypeScript — REST integration (typed fetch)
const res = await fetch("https://api.biomedora.com/v1/reasoning/query", {
method: "POST",
headers: {
Authorization: `Bearer ${process.env.BIOMEDORA_TOKEN}`,
"Content-Type": "application/json",
"Idempotency-Key": crypto.randomUUID(),
},
body: JSON.stringify({
patient_id: "patient_123",
question: "What evidence suggests renal deterioration?",
}),
});
const answer = await res.json(); // claims[], citations[], uncertaintySecurity & Privacy
Protect sensitive data at every step
Privacy controls start at upload and continue through processing, AI use, release, and audit.
Full security architectureRelease only when it is safe
Detect PHI, create stable pseudonyms, shift each patient’s dates consistently, and scan again. If PHI remains, the document is blocked.
Keep every organization separate
The server sets the tenant on every request. Cross-tenant access is denied and audited—even for administrators.
Control where AI runs
Use deterministic processing first and local AI by default. External AI requires minimized PHI and explicit approval, and can be disabled completely.
Prove what happened
Every sensitive action records who, what, when, outcome, and request ID in an append-only, tamper-evident audit trail.
Choose where it runs
Use a single-tenant cloud, your own AWS account, or an on-premises deployment with no external data egress.
Treat documents as evidence—not instructions
Hidden instructions inside a document are surfaced for review and never passed to a model as commands.
Start with one hard document problem
Turn it into a workflow your team can trust.
Bring pathology reports, patient charts, trial documents, safety narratives, genomic reports, or a product idea. We’ll map the right BioMedora workflow and deployment option.
Explore · 1,000 pages
Build · $49/month
Collaborate · 25,000 pages
BYOC and on-premises