Blue-IQ

Any document in.
Structured data out, scoredfield by field.

Resumes, contracts, invoices, licences. Capture pulls out the fields that matter and tells you how sure it is about each one, so your team stops re-typing and only checks what is actually in doubt.

Book a demoRead the docs

Core Features

Fast where it can be. Careful where it counts.

Three things every document team asks for,
and most parsers quietly skip

Registered Nurse with 6 years on a 32-bed telemetry unit, holding a compact TN licence and BLS certification
✦Review low scores

Field-level confidence

API access

Search every parse

Parsed record library

Paperwork in. Payload out. Three stages.

  1. 01

    Ingest anything

    PDFs, scans, exports, whatever landed in the inbox. Post them to the API or drop them in a watched folder. If someone photographed it on a phone, OCR picks it up.

    PDFDOCXScan
    api - watched folder - ocr
  2. 02

    Read and score with Sonar

    Sonar reads the document and scores itself as it goes. Where it is unsure, it says so. Where it has nothing, it returns null. It will not invent a value to fill a gap.

    profession1.00
    specialty0.94
    facility0.88
    confidence: { field: 0.91 }
  3. 03

    Deliver where you work

    Validated JSON goes straight into your ATS, CRM or warehouse over a documented REST API and signed webhooks. Nobody re-types anything.

    ATSCRMWarehouse
    -> ATS - CRM - warehouse
POST /api/v1/parsefull reference ->
# one call, any document
curl -X POST \
  https://api.parsinglab.blue-iq.ai/api/v1/parse \
  -H "X-API-Key: cap_live_..." \
  -F "file=@contract.pdf"
// 200 OK
{
  "status": "completed",
  "data": { "parties", "term", "total"... },
  "confidence": { "overall": 0.94 },
  "fabricated": false
}

Most parsers guess. This one admits what it does not know.

A generic model flattens a credential into word salad and fills the gaps with something plausible. Capture reads the documents your business actually runs on, and puts a number on its own certainty.

profession1.00
RN
specialty0.94
Med Surg / Tele
facility0.82
Fort Sanders Regional

Confidence on every field

Every value carries a number. Review the weak ones, ship the rest.

stateTN
licence_nonull
expiresnull
2 flagged for review

Never fabricates

No value beats a wrong value. Unsure fields come back null, flagged.

threshold0.85
12 to review488 auto

Human-in-the-loop by design

Pick a threshold. Anything under it goes to a person. Everything over it just goes.

{
  "profession": "RN",
  "specialty_id": "88",
  "confidence": 0.94
}
Schema valid

Schema-validated output

JSON that matches your schema on arrival. Not a wall of text to untangle later.

These are real people's records. We treat them that way.

Clinical histories, signed agreements, salary lines. What you send us is not test data, and the handling reflects that.

TLS 1.3

api.parsinglab.blue-iq.ai

X-API-Key: rp_live_••••••••

Encryption in transit

TLS on every request, authenticated by a per-workspace key you can rotate whenever you like.

acme

globex

no shared access

Workspace isolation

Documents and keys never leave their workspace. Role-based access and SSO on top.

zero_retention
parsed in memory
returned to caller
discarded

Zero-retention option

Switch on zero retention and the document is parsed in memory and gone. Nothing stored is nothing to leak.

duration22.7s
file_typepdf
tokens38,905
content
never logged

Content-free audit trail

We log that a parse happened, how long it took and what it cost. Never a word of what it said.

Live, on your files

Send us your worst document.

The messy scan. The seven-page contract. The resume with three job titles crammed into one line. We will run it live and show you exactly what comes back.

Talk to usRead the docs
contract.pdfcompleted
counterparty0.96
Fort Sanders Regional
governing_law0.91
Tennessee
term_end0.88
2027-06-30
auto_renew0.00
null

1 field flagged - nothing invented