Document intelligence · DPDP-native

Every form.
Every script.
Structured.

DoqSeal turns India's most chaotic paper trail — handwritten ledgers, dilapidated government forms, scanned PDFs in nine languages — into clean, routable JSON.

DPDP Act · Section 11 compliant SOC 2 Type II Data residency · India
Indic languages9+
Field accuracy98.4%
Median latency1.2s
api.doqseal.com / extractlab_report_4421.pdf
VERIFIED · ०१
पैथोलॉजी रिपोर्ट · LAB REPORT
Patient IDAP-4421-B
मरीज़ / Patientप्रिया शर्मा
TestComplete Blood Count
LabApollo Diagnostics
Ref. DoctorDr. Mehta
ParameterValueRange
Hb (g/dL)12.812.0–15.5
WBC (×10³)7.24.0–11.0
Platelets248150–410
Date18 Mar 2025
PathologistDr. N. Rao
extract.json LIVE
{
"document_type": "diagnostic_report",
"patient_name": "प्रिया शर्मा",
"test_type": "complete_blood_count",
"hemoglobin_gdl": 12.8,
"lab": "Apollo Diagnostics",
"report_date": "2025-03-18",
"language": "hi-IN, en-IN",
"confidence": 0.987,
"verified": true
}
The Platform

From pixels to production, in one pipeline.

No middleware. No manual review queue. No exporting CSVs at midnight. Documents land, DoqSeal structures them, your systems wake up to clean rows.

01 · Capture

Ingest anything paper-shaped.

PDFs, scans, mobile photos, faxes, email attachments. Even forms photographed in poor light at a tehsildar's office.

khata_287.pdfscan_3.jpg · noisyinvoice_8892.pdfemail.emlform_60.tiffmedical.png
02 · Extract

Vision models trained on Indian paperwork.

Proprietary neural extraction reads degraded scripts, faded carbon copies, mixed-language forms — and emits typed JSON.

"name": "P. Singh"
"village": "Moga"
"amount": 284500
"date": "2025-12-12"
"verified": true
03 · Route

Straight into the systems you already run.

Data flows directly to Salesforce, Tally, SAP, your custom ERP — over typed APIs and webhooks. Manual entry, retired.

→ salesforce.opportunity200 OK
→ tally.voucher.create200 OK
→ webhook · audit.log200 OK
→ s3 · vault/encrypt200 OK
Regional OCR

The first OCR engine fluent in dilapidated.

Most engines hallucinate on a faded panchayat ledger. DoqSeal was trained on 14 million pages of real Indian paperwork — handwritten, photocopied, hole-punched, water-stained. Try it on a script.

14M
Pages trained
98.4%
Field accuracy
9+
Indic scripts
बैंक स्टेटमेंटHI-IN · scan_hi_002.pdf
खाता संख्या ४००२-१९८३ के अंतर्गत कुल जमा राशि ₹४,७५,३२० है। लेन-देन की अंतिम तिथि १५ अक्टूबर २०२५।
extract.json98.7% confidence
"account_no": "4002-1983",
"balance_inr": 475320,
"last_txn_date": "2025-10-15",
"language": "hi-IN",
"confidence": 0.987,
Vault Architecture

We process what we can't read.

The DoqSeal Vault is a zero-knowledge processing layer. Documents are encrypted on your side with AES-256. We store only ciphertext.

PLAINTEXT · CLIENTAES-256
CIPHERTEXT · DOQSEAL0xE4...
1eA1 F72e 5Dcf 3d53 C4Ca 12Ef D9B2 cB71 2f73 C338 5ff6 FeE4 fc2d Fd4C DdED DACC 3a4b c84d 5ff6 4bbb AC6c 8B48 8eb0 0C1F FE6e b65A faEA D9F9 e05d 7Eb5 33a7 210d CEdE 0E45 82a0 dE2c ED3C 3bDe 5Fcc d5E5 85F7 CEeB f62A 4bfD aAb6
key held by client
CLIENT-SIDE
AES-256-GCM
KEY CUSTODY
Yours, always
ENCLAVE
Nitro / TEE
RESIDENCY
Mumbai · Hyderabad

Extraction happens inside an enclave where keys exist for milliseconds, then evaporate. We never persist plaintext. We never can.

Compliance

DPDP-native. Not retrofitted.

Most platforms bolt on compliance the quarter before audit. DoqSeal was architected after the Digital Personal Data Protection Act passed — every component carries a clause it answers to.

DPDP Section 8

Data localization

All processing on servers physically located within India. No cross-border transfer without explicit consent.

MEITY-CERTIFIED REGION
DPDP Section 11

Right to erasure

Cryptographically verifiable deletion across primary, replica and audit stores within 72 hours of request.

T+72H GUARANTEE
DPDP Section 10

Immutable audit log

Every read, write and access event hash-chained and timestamped. Tamper-evident by design.

SHA-256 CHAINED
ISO 27001

Vault architecture

AES-256 client-side encryption. We hold ciphertext only — encryption keys never leave your perimeter.

ZERO-KNOWLEDGE
RBI

Financial-grade controls

Aligned with RBI master directions on outsourcing, IT governance and information security for regulated entities.

FOR REGULATED ENTITIES
SOC 2 II

Independent attestation

Type II report covering security, availability, confidentiality. Renewed annually, available under NDA.

RENEWED 2025
DPDP Section 6

Granular consent

Per-document, per-field, per-purpose consent receipts. Revocable at any time, propagated downstream.

CONSENT-AWARE
HIPAA

Health data ready

Optional HIPAA-aligned controls for medical records, prescriptions and diagnostic reports.

OPTIONAL MODULE
Use cases

Six verticals. One platform.

From an NGO reconciling 80G receipts to a State Data Centre digitising thirty years of land records — DoqSeal handles the document where it lives, in the language it was written.

Integrations

Your CRM. Your ERP. Your warehouse.

SalesforceCRM
TallyAccounting
SAPERP
ZohoSuite
HubSpotCRM
SalesforceCRM
TallyAccounting
SAPERP
ZohoSuite
HubSpotCRM
SalesforceCRM
TallyAccounting
SAPERP
ZohoSuite
HubSpotCRM
SalesforceCRM
TallyAccounting
SAPERP
ZohoSuite
HubSpotCRM
Developers

Five lines, and you're extracting.

No SDK lock-in. No webhook rituals. Authenticate once, POST a document, receive typed JSON with field-level confidence scores. Ship to production same week.

ENDPOINT
POST api.doqseal.com/v1/extract
MEDIAN LATENCY
1.2s · single-page document
PRICING
₹0.40 · per page · pay-as-you-go
cURLPythonNode
v1 · stable
# Extract a multi-language land record
curl https://api.doqseal.com/v1/extract \
  -H "Authorization: Bearer $DOQSEAL_KEY" \
  -H "X-Residency: in-mumbai" \
  -F "file=@khata_287.pdf" \
  -F "schema=land_record_v3" \
  -F "languages=hi,ta,en"

Replace data entry with data engineering.

500 pages free. No credit card. Production-ready API in under five minutes.