M37Labs Logo
M37Labs Pehchaan Sovereign Aadhaar KYC Model
IDENTITY SLM · BILINGUAL REGIONAL SCRIPT EXTRACTION · ZERO EGRESS

Sovereign Aadhaar verification with zero cloud egress.

Purpose-built for Indian identity KYC at scale. Offline UIDAI RSA 2048-bit Secure QR verification, regional Indic language extraction, autonomous 8-digit masking, and granular 9-part address decomposition.

01/Core Fields Pipeline
95.2%Name, DOB, Gender, Aadhaar No.
02/Visual Fallback
85.7%When QR is unverified or torn
03/Critical Invariants
100%DOB, Gender, Aadhaar and VID match
04/Secure QR Specs
V1-V5UIDAI RSA 2048-bit validation
05/Address Schema
9 PartsDiscrete structural decomposition
06/Regulatory Masking
DefaultAutomatic 8-digit redaction
THE IDENTITY COMPLIANCE MANDATE

Processing Aadhaar at scale requires cryptographic verification and compliance.

Aadhaar is the identity backbone for over 1.4 billion citizens. Modality fragmentation (PVC cards, laminated paper, mobile captures, e-Aadhaar PDFs) and strict regulatory masking mandates make cloud OCR pipelines unreliable and non-compliant. Pehchaan provides deterministic offline verification, bilingual regional script parsing, and automatic 8-digit PII redaction.

01 / CRYPTOGRAPHY

Offline UIDAI RSA 2048 Validation

Validates the tamper-proof RSA 2048-bit digital signature embedded in UIDAI Secure QR codes (V1 through V5) completely offline without making external calls to government endpoints or risking network downtime.

V1 through V5 Secure QR Specifications
02 / COMPLIANCE

Autonomous 8-Digit Redaction

Automatically masks the initial 8 digits of the 12-digit Aadhaar number across all extracted JSON payloads and spatial bounding coordinates, ensuring instant compliance with RBI and DPDP statutory frameworks.

100% Extraction Rate on Core Invariants
03 / NORMALIZATION

9-Part Address Structural Split

Disassembles unstructured multilingual Indic address blocks into standardized entities: house, street, locality, landmark, VTC, sub-district, district, state, and PIN code, eliminating manual verification overhead.

Regional Indic Languages Supported
CORE ARCHITECTURAL ADVANTAGES

Why Pehchaan? Five Core Innovations.

Engineered specifically to solve real-world KYC failure modes, linguistic diversity, and regulatory PII compliance.

RSA 2048 VERIFIED

UIDAI Secure QR (V1-V5) Verification

Natively parses all five generations of UIDAI QR codes, from legacy uncompressed text to modern 2048-bit RSA-signed gzip binary streams. Validates signatures completely offline against official UIDAI public key certificates.

Offline UIDAI Root Key MathValid RSA Signature
Mathematical ProofZero external API dependencies
REGIONAL INDIC SCRIPTS

Bilingual Name & Script Extraction

Simultaneously parses cardholder names and demographic details in both English and regional scripts (Devanagari, Tamil, Telugu, Kannada, Bengali, Gujarati, Marathi, Malayalam, and Punjabi, among others) without phonetic hallucinations.

Dual Hindi/Regional and Latin phonetic cross-verification prevents spelling mismatch drop-offs
Bilingual Token HeadNative Indian language coverage
9 ENTITIES

9-Part Address Decomposition

Intelligently decomposes messy unstructured Indian addresses into standardized schema: house, street, locality, landmark, VTC, sub-district, district, state, and PIN code.

Structured Address Output
DPDP & RBI

Autonomous 8-Digit Masking

Automatically redacts the first eight digits of the Aadhaar number across all extracted JSON metadata and spatial bounding boxes to support statutory masking requirements.

Redacted Before In-Memory Storage
95.2% OVERALL

Physical Degradation Resilience

Maintains 85.7% accuracy even when physical cards are stained, folded, or have torn QR codes, using vision-language fallback to complete customer onboarding.

100% Extraction on Critical Invariants
INTERACTIVE VERIFICATION SUITE

Live Scenario Inspector & Telemetry

Test bilingual parsing, UIDAI RSA Secure QR validation, default Aadhaar masking, and granular 9-part address decomposition.

DOCUMENT SURFACE
GOVERNMENT OF INDIA
भारत सरकार
BILINGUAL NAME
AARAV SURESH KULKARNI
आरव सुरेश कुलकर्णी
DOB / जन्म तारीख21/04/1991
GENDER / लिंगMALE / पुरुष
AADHAAR NUMBER
XXXX XXXX 7892
VID: 9102 3847 1290 4481
CORE ACCURACY98.7%
UIDAI RSA SIGNATUREVERIFIED (RSA 2048)
SECURE QR VERSIONV5 (Secure 2048-bit)
DPDP COMPLIANCE100% Masked
response.json
{
  "status": "SUCCESS",
  "document_type": "AADHAAR_PVC_CARD",
  "verdict": "GENUINE",
  "genuineness_score": 0.987,
  "demographics": {
    "name_english": "AARAV SURESH KULKARNI",
    "name_regional": "आरव सुरेश कुलकर्णी",
    "regional_language": "marathi",
    "date_of_birth": "1991-04-21",
    "gender": "MALE",
    "care_of": "S/O Suresh Kulkarni"
  },
  "identifiers": {
    "aadhaar_number_masked": "XXXX-XXXX-7892",
    "virtual_id": "9102384712904481"
  },
  "invariants_status": {
    "dob_extracted": true,
    "gender_extracted": true,
    "aadhaar_extracted": true,
    "vid_extracted": true,
    "invariant_accuracy": "100%"
  },
  "cryptographic_verification": {
    "qr_detected": true,
    "qr_generation": "V5",
    "rsa_2048_signature_valid": true,
    "public_key_authority": "UIDAI_ROOT_CA_2023",
    "decompression_algorithm": "GZIP_STREAM"
  }
}
M37Labs Enterprise AI Computing Lab
100% BEHIND FIREWALL · ZERO DATA EGRESS · BUILT FOR RBI & DPDP

Sovereign Aadhaar verification at enterprise scale.

95.2% accuracy. Offline UIDAI RSA 2048 verification. Automated 8-digit masking. Deploy Pehchaan directly inside your secure VPC, banking servers, or branch kiosks without exposing customer PII to public clouds.