Find the ICD-10 codes your documentation supports but your claims are missing.
Chart audit API that surfaces missed HCCs with RAF-weighted revenue estimates, unsupported codes for RADV defense, specificity upgrades, and denial risk. Every finding carries the exact text span that supports it.
Turn clinical notes into ICD-10 Codes in seconds, too.
Every chart has money hiding in it.
Five finding types, one endpoint, evidence spans on everything. Built on CMS PY2026 v22 + v28, WHO ICF, and SNOMED/UMLS data.
HCC Gap Capture
$9,016
avg per CHF gap
Missed HCCs with RAF × base-rate revenue estimates, emitted once per v22/v28 model when a code is HCC in both.
RADV Defense
$4,198
avg exposure per unsupported
Codes with no documentation support, each with a reason and what would have supported it.
Specificity Upgrades
E11.9 → E11.311
unspecified → precise
When the text names a distinguishing feature the submitted code doesn't carry, propose the specific child.
Denial Risk
High / Moderate / Low
documentation-quality tiers
Flags missing laterality, missing episode of care, age/sex mismatches, unspecified-vs-subtype gaps.
Problem List
Active vs Historical
deduplicated, evidence-linked
Reconciles all active conditions from the documentation, dedupes by ICD-10, respects negation and history flags.
How It Works
Go from clinical notes to coded claims in three simple steps.
Send Clinical Text
Paste notes into the playground or call the API. Progress notes, discharge summaries, consult reports. Any format works.
AI Extracts Diagnoses
Medical NLP identifies conditions, rules out negated findings, and resolves abbreviations and clinical context.
Get ICD-10 Codes
Receive accurate ICD-10-CM codes with confidence scores, ready for your EHR, billing system, or automated workflow.
Send Clinical Text
Paste notes into the playground or call the API. Progress notes, discharge summaries, consult reports. Any format works.
AI Extracts Diagnoses
Medical NLP identifies conditions, rules out negated findings, and resolves abbreviations and clinical context.
Get ICD-10 Codes
Receive accurate ICD-10-CM codes with confidence scores, ready for your EHR, billing system, or automated workflow.
Developer? Here's how simple it is.
TypeScript SDK
// npm install autoicd-js
import { AutoICD } from "autoicd-js";
const client = new AutoICD({ apiKey: "sk_..." });
const result = await client.code(
"Type 2 diabetes with chronic kidney disease stage 3"
);Response
{
"entities": [{
"entity_text": "Type 2 diabetes with chronic kidney disease stage 3",
"codes": [{
"code": "E11.22",
"description": "Type 2 DM with diabetic CKD",
"confidence": "high"
}, {
"code": "N18.30",
"description": "CKD, stage 3 unspecified",
"confidence": "high"
}]
}]
}See It in Action
Real clinical text in, structured ICD-10 codes out.
Extracted Entities
Free to start. No credit card required.
Built for Real Clinical Documentation
From messy progress notes to billable ICD-10-CM codes, with negation detection, abbreviation expansion, and confidence scoring.
AI-Powered Accuracy
Specialized medical AI reads your clinical text, identifies every diagnosis, and matches it to the right ICD-10 code, just like an experienced coder.
74,000+ ICD-10 Codes
The complete 2025 ICD-10-CM code set with expanded synonyms and abbreviations so even shorthand notes get coded correctly.
Full Inter-Standard Mapping
Every code includes cross-references across ICD-10, ICD-11, ICF, LOINC, SNOMED CT, and UMLS for complete clinical data interoperability.
Rules Out Negated Findings
Understands phrases like "no evidence of heart failure" and only codes conditions that are actually present.
Understands Clinical Context
Distinguishes between current diagnoses, past medical history, family history, and uncertain findings for precise coding.
PHI De-identification
Automatically detect and mask patient names, dates, SSNs, phone numbers, and other protected health information before it leaves your system.
HIPAA Compliant
Patient data is processed in memory only. Never stored, logged, or used for training. BAA available for Pro users.
ICF Functioning Codes
Map diagnoses to WHO ICF categories for disability assessment, rehabilitation planning, and functional evaluation. Built-in Core Sets for 12+ conditions.
Integrate Anywhere
TypeScript and Python SDKs, an MCP server for Claude and Cursor, and a simple REST API.
Simple, Transparent Pricing
7-day free trial, no credit card required. Upgrade when you're ready.
$0
- 100 coding requests/day
- All API endpoints
- Interactive playground
- PHI de-identification
- 7-day trial
$49/mo
- 1,000 coding requests/day
- All API endpoints
- Interactive playground
- PHI de-identification
- Email support
$199/mo
- 10,000 coding requests/day
- All API endpoints
- Interactive playground
- PHI de-identification
- BAA available
- Priority support
Contact us
- Unlimited requests
- Dedicated support
- Custom BAA & SLA
- On-premise deployment available
All plans include the interactive playground. Trial requires no credit card.
Frequently Asked Questions
What is automated ICD-10 coding?
Automated ICD-10 coding uses AI to read clinical text (progress notes, discharge summaries, etc.) and assign the correct ICD-10-CM diagnosis codes. Instead of manually searching through 74,000+ codes, the software identifies conditions in the text and matches them to codes in seconds.
How accurate is the coding?
AutoICD uses a purpose-built pipeline of custom-trained clinical NLP models — not a generic LLM wrapper. Every code returned is a real, valid ICD-10-CM code with a confidence score so you can audit edge cases. Unlike prompt-based approaches that hallucinate non-existent codes, our deterministic pipeline produces consistent results: same input, same output, every time. Negated findings like "no evidence of heart failure" are automatically excluded.
Is my patient data safe?
Yes. AutoICD API is fully HIPAA compliant. Patient data is processed in memory only and is never stored, logged, or used for model training. All data is encrypted in transit with TLS. Business Associate Agreements (BAAs) are available for Pro users. See our security page for full architecture details.
Which ICD-10 codes are supported?
AutoICD API includes the complete 2025 ICD-10-CM code set with over 74,000 diagnosis codes and expanded synonyms and abbreviations for better matching accuracy.
Does the API include cross-standard mappings?
Yes. Every code lookup and coding response includes cross-references across ICD-10, ICD-11, ICF, SNOMED CT, and UMLS. The coding endpoint returns mapped ICD-11 codes, SNOMED CT concept IDs, UMLS CUIs, and related ICF categories for each match.
Does AutoICD support ICD-11 codes?
Yes. AutoICD includes the complete ICD-11 MMS code set with over 17,000 codes, WHO index term synonyms, SNOMED CT and UMLS cross-references, and bidirectional ICD-10 ↔ ICD-11 mapping. Use the crosswalk converter to look up equivalent codes between versions using WHO's official mapping tables.
What is ICF coding?
ICF (International Classification of Functioning, Disability and Health) is a WHO classification that describes how health conditions affect a person's functioning — body functions, activities, participation, and environmental factors. AutoICD maps clinical text to ICF categories automatically, complementing ICD-10 diagnosis codes with functioning assessment data used in rehabilitation, disability evaluation, and care planning.
What are ICF Core Sets?
ICF Core Sets are curated shortlists of ICF categories relevant to specific conditions (e.g., diabetes, stroke, depression). AutoICD includes built-in Core Sets for 12+ conditions, so given an ICD-10 diagnosis code, you can instantly get the relevant ICF functioning categories to assess.
Ready to Automate Your Medical Coding?
Start with 100 free coding requests. No credit card, no setup.