Core asset · Foundation428 roles · 175K aliases

Taxonomy Mapper.

"Med-Surg RN", "MS Registered Nurse", "RN — Medical Surgical" — three records, one canonical role. Your messy data, reconciled before it enters a distribution.

Book a data demo →See the anatomy

428Canonical roles

175,231Aliases mapped

≥ 0.92Confidence floor

Inline · API · BulkResolution

Anatomy By the numbers Capabilities Endpoint Under the hood Use cases What's included FAQ

Anatomy

What the mapper resolves, top to bottom.

01Raw title in

Whatever your ATS, CRM, or VMS calls it — including punctuation, abbreviations, and contract-length suffixes.

02Canonical role out

One of 428 entries. Every chart, alert, and API response across the platform speaks this language.

03Confidence score

0.00 — 1.00. The platform refuses to resolve below 0.92 and surfaces the ambiguity for human review.

04Override path

Teams can pin overrides for proprietary titles. Pinned mappings flow back into every downstream surface.

By the numbers

The reconciliation layer behind the platform.

428Canonical rolesTravel, per diem, local contract — separate cohorts.

175KAliases mappedCurated from production ATS / VMS / contract data.

0.92Confidence floorRecords below threshold route to human review, never silently bucketed.

100KBulk normalize / callEnterprise endpoint capacity, single batch.

Capability matrix

What Taxonomy Mapper ships with.

Capability	Desk · Agency · Agency Plus	Enterprise
Canonical taxonomy access	Read · suggest	Read · suggest · override
Bulk normalize endpoint	—	100K rows / call
Override pinning	Per-user	Org-wide · audited
Confidence reporting	Per-row	Per-row · per-source
Custom taxonomy extensions	—	Per-tenant namespace

Endpoint

One call. One canonical answer.

$ curl "https://api.clinicalrate.com/v1/normalize" -H "Authorization: Bearer $CR_TOKEN" \
    --data '{"title": "MS RN — Travel 13wk"}'
# 200 OK
{
  "input": "MS RN — Travel 13wk",
  "role_id": "CR-0142",
  "canonical": "Travel RN — Med/Surg",
  "confidence": 0.98,
  "contract_class": "travel-13wk",
  "alternates": [
    { "role_id": "CR-0148", "canonical": "Travel RN — Telemetry", "confidence": 0.04 }
  ]
}

Under the hood

How an alias becomes canonical.

LexicalAbbreviation expansion + token stemming

"MS", "M/S", "Med-Surg", "MedSurg" — all normalized to a common form before retrieval. Built on a curated clinical abbreviation set, not generic NLP.

SemanticEmbedding retrieval across 175K aliases

Top-k similarity over the canonical alias corpus. The embedding model is fine-tuned on clinical role data — generic LLMs confuse "RT" (Respiratory Therapist) with "Radiology Tech".

DisambiguationContract-class + specialty splitter

"Travel" vs "per diem" vs "local contract" extracted as a separate dimension. ICU vs Med/Surg vs L&D never share a distribution by accident.

AuditEvery override is logged

When an operator pins an override, the change is timestamped, attributed, and reversible. The audit log is exportable to your compliance stack on Enterprise.

Enterprise tier · bulk normalize endpoint

Priya ShahHead of Data · Hospital system, 14 facilities

“We deleted three pages of regex from our ETL the week we plugged this in. Specialty distributions stopped being a debate and started being a benchmark.”

Where it shows up

Three teams, three reasons to call /normalize.

Agencies

Inline at quote time

Resolve raw req titles in your ATS to canonical roles before pricing — no more "what specialty is this?" tickets.

0.98Median confidence Hospitals & facilities

Clean reporting

Joinable canonical role IDs across your VMS, payroll, and contracts — one specialty, one row.

14Source systems consolidated AI systems

Reasoning agents

A canonical role ID is the join key your pricing agent uses against rates, alerts, and trends — without LLM hallucinations.

<40 msp50 normalize latency

What's included

Taxonomy Mapper by tier.

ClinicalRate · All plans

From $1,000/ month

Canonical taxonomy and per-user overrides on every plan. The /normalize API, bulk endpoint, and per-tenant extensions come with Enterprise.

Read access to all 428 canonical roles
/normalize endpoint with confidence (Enterprise)
Bulk normalize up to 100K rows / call (Enterprise)
Per-row contract-class disambiguation
Override pinning (per-user · org-wide on Ent.)
Per-row and per-source confidence reporting
Audit log exportable to compliance stack
Per-tenant namespace extensions (Enterprise)

Book a data demo →See full pricing

FAQ

Questions buyers ask about Taxonomy Mapper.

What happens to titles you can't resolve?

Anything below the 0.92 confidence floor is routed to a human-review queue with the top three candidate canonicals and their scores. Reviewed mappings are added to the alias corpus on the next nightly rebuild.

Can we keep proprietary internal titles?

Yes. Every plan supports per-user pinned overrides; Enterprise adds org-wide, audited overrides. Pinned mappings flow through every downstream surface (Rate Explorer, Alert Stream, exports) without altering the global canonical taxonomy.

How do you keep the canonical list current?

The canonical taxonomy is versioned. New roles enter through a quarterly review process; we publish a changelog and a migration note for any non-additive change. Pinned overrides are unaffected.

Is this an LLM under the hood?

A clinical-domain embedding model handles the semantic step, but lexical expansion, contract-class extraction, and the override layer are deterministic. Confidence scores are calibrated on a held-out clinical corpus, not a generic benchmark.

Other features

The rest of the surface.

Get started

Reconcile your data to one taxonomy.

Book a demo and bring a sample of your messiest role titles. We'll run them through live.

Book a data demo →