Medical coding infrastructure, behind one API.

The self-service platform from HANK.ai: code sets, claim scrubbing, de-identification, and automated coding under one subscription, with usage across all of it. Every product ships a web app, a REST API, and an MCP endpoint. Built by medical coding and billing experts and their engineers: the same engines that underpin HANK.ai's enterprise systems.

Get started Read the docs

api.hank.ai

$ curl -H "Authorization: Bearer $HANK_API_KEY" \
  -d '{"model": "hank/fast", "messages": [{"role": "user",
  "content": "Can 99213 be billed with 20610 same visit?"}]}' \
  https://api.hank.ai/v1/scholar/hank/v1/chat/completions

"Yes, but only if the 99213 is a significant, separately
 identifiable E/M service: bill 99213-25 with 20610 on the
 same date. A routine visit for the injection alone should
 not get a separate E/M." [1][2]

"citations": [
  { "id": 1,
    "document_title": "Modifier -25: Significant,
        Separately Identifiable E/M Service",
    "section": "E. Multiple E/M Services",
    "snippet": "...Significant, Separately Identifiable
        E/M Service by the Same Physician... on the
        Same Day of the Procedure...",
    "open_url": "modahealth.com/.../RPM028_Mod25SigSepEM.pdf" },
  { "id": 2,
    "document_title": "HANK Codesets: CPT 20610",
    "codesets_profile": { "code": "20610", "is_valid": true },
    "open_url": "codesets.hank.ai/lookup?q=20610" } ]

Products

Nine services. One subscription, one key, one console.

Codesets

The unified U.S. medical-coding reference: look up, validate, and crosswalk the major procedure, diagnosis, drug, and provider code sets, plus HCC risk scoring, claim groupers, NCCI edits, and fee schedules, all date-of-service aware.

codesets.hank.ai

ClaimCleaner

Validates whole claims and every line item on them: codes, modifiers, units, and diagnoses checked against CMS billing edits, NCCI, MUE, LCD coverage, and fee schedules, with suggested corrections and a full audit trail. Sub-second per claim.

claimcleaner.hank.ai

Scholar

Sourced medical-coding answers from a trusted, validated knowledge base. Every answer cites its source, with authoritative code descriptors from Codesets.

scholar.hank.ai

ARSE (Automated Recognition and Structured Extraction)

Ingests inbound payer mail, EOBs, and correspondence as fax, scan, PDF, or image. Vision models extract every patient, document, and claim, and emit gated X12 835 remittances for clean EOBs.

arse.hank.ai

Payer Policy Manager

Monitors hundreds of U.S. payer-policy portals for changes, summarizes each update, and extracts its billing rules. Alerts by API, webhook, RSS, or email, plus a hospital price-transparency database.

ppman.hank.ai

Automated Medical Coding

Turns clinical notes and PDFs into structured coded claims: procedure codes, modifiers, diagnoses, and review flags, with reflexive coding and bundling built in.

krac.hank.ai

SafeHarbor

HIPAA Safe Harbor de-identification for clinical notes and PDFs: detects the 18 identifiers and replaces each with a consistent, realistic fake instead of a redacted hole, so documents stay readable.

safeharbor.hank.ai

MedicalNotes

Submit a clinical record as text or PDF: get a chart summary, a structured note of your chosen type, or classification by note type and date of service.

medicalnotes.hank.ai

Docs

Documentation for every HANK.ai service, written for humans and language models alike: guides, API references, changelogs, full-text and semantic search, and Ask AI answers.

newdocs.hank.ai

Platform

Boring on purpose, where it counts.

Web, API, and MCP

Every product has a full web app, a REST API at api.hank.ai/v1/<product> with a bearer key you mint in the console, and an MCP endpoint your agents can call. Same gateway, same key, all three.

One subscription, metered usage

Subscribe once, use every product. Usage is metered per call and visible in the console as it happens. Set per-key rate and volume limits yourself. No surprise invoices.

Built for PHI, gated by a BAA

The clinical services (Automated Medical Coding, MedicalNotes, SafeHarbor, ARSE) are built to process patient data. On qualifying plans, HANK.ai signs a Business Associate Agreement, and PHI flows only after it is executed. Without a BAA the platform is PHI-free by policy. Ask from your console account.

Operated in the open

Live service health at /status, versioned docs at newdocs.hank.ai, and a real changelog. When something breaks, the status page says so.

Agents

Bring your agent. It can set everything up.

Paste this into Claude, Codex, or any agent

The whole platform is agent-readable: llms.txt discovery, per-service MCP servers, per-page Markdown docs. Your agent reads the docs, walks you through sign-in and an API key, connects itself, and makes the first call.

Full guide: console.hank.ai/console/docs/agent-setup

Read https://api.hank.ai/llms.txt to learn the HANK platform: every service is a REST API and an MCP server behind one bearer token. Then get me set up: (1) I sign in at https://console.hank.ai/signin with Google. (2) Help me mint an API key at https://console.hank.ai/console/api-keys and scope it to only the services I need. (3) Register the MCP servers I want (https://api.hank.ai/v1/<service>/mcp with header Authorization: Bearer <my key>) or call the REST APIs at https://api.hank.ai/v1/<service>/. (4) Docs are at https://newdocs.hank.ai (llms.txt and per-page Markdown available). Make one small test call, like asking HANK Scholar a coding question, then summarize what you can do for me.

CPT® is a registered trademark of the American Medical Association. Code-set content is subject to its owners' licenses; see the Terms of Service for the licenses you must hold.