Back to Products
Product · Distil CV Intelligence
Live · Invite-only betaUpload a CV. Get structured data. In seconds.
Distil extracts 20+ fields from any CV — PDF, DOCX, or HTML — maps them to NCS and UGC standards for India, scores confidence per field, and flags only what needs a human eye.
20+
Fields Extracted
6-step
AI Pipeline
NCS + UGC
India Standards
< 2s
Per CV
Sector
HR Tech · Recruitment
Standards
NCS · UGC · ESCO · ISCED
Status
Live · Invite-only beta
Input formats
PDF · DOCX · HTML
The Problem
Manual CV screening doesn't scale.
Manual Screening
- HR reads every CV manually — hours per batch
- Data copy-pasted into spreadsheets — errors everywhere
- Qualification levels judged inconsistently per recruiter
- No structured output — no way to search or filter at scale
- Duplicate candidates — no deduplication logic
- No audit trail on who reviewed what and when
With Distil
- Upload CVs in bulk — structured data extracted in seconds
- Every field parsed, normalized, and scored automatically
- Qualifications mapped to UGC levels — consistent every time
- Search and filter candidates by any structured field
- Flagged fields routed to human review — nothing missed
- Full audit log — every action timestamped and tracked
How It Works
A 6-step pipeline — from raw file to structured record.
Every CV goes through the same deterministic pipeline. Each step is tracked and visible in the UI.
01Ingestion
Upload PDF, DOCX, or HTML. File stored securely in Supabase Storage. Job queued instantly.
02Document Parsing
Native text extracted from PDFs and DOCX files using pdf-parse and Mammoth. Image-heavy files flagged for OCR.
03OCR Fallback
Scanned or image-based CVs are detected automatically. Native text returned when available.
04LLM Extraction
Groq Llama 70B extracts 20+ structured fields — name, email, phone, location, experience, education, skills, certifications, and salary expectation. Indian CV awareness built in (education at the bottom, 8000-char window).
05Standardization
Occupations mapped to NCS codes (India) or ESCO codes (EU). Education mapped to UGC levels (India) or ISCED levels (EU). Dates normalized to ISO 8601. Phone to E.164.
06Validation & Scoring
Per-field confidence scored. Weak or missing fields flagged for human review. Final accuracy percentage computed and stored.
Features
Built for real recruitment workflows.
Indian Degree Intelligence
Handles every Indian degree — B.Tech, MBA, MIB, B.Com, CA, CS, CMA, BHM, B.Des, MBBS, and 30+ more. Rule-based UGC level assignment works for any degree, listed or unlisted. Highest qualification auto-detected.
Human Review Interface
Flagged fields shown to the reviewer one by one. Accept or reject each extraction. Overrides stored in the audit log. Review only what the AI wasn't confident about.
Reprocess Anytime
Re-run the full AI pipeline on any already-uploaded CV. Useful after prompt improvements or when a CV was processed with older models. Original file never deleted.
Invite-only Access
Access is gated — superadmin invites specific email addresses. Invited users receive a branded email and can sign in. No open sign-ups.
Audit Trail
Every upload, extraction, field override, and export is logged with timestamp and user. CSV export available for compliance and audit purposes.
Approve & Export
Once reviewed, approve the candidate record and export the structured data. Integrations and webhooks available for pushing to ATS, HRIS, or job portals.
Standards Compliance
Structured to globally recognized standards.
Every extracted record is mapped to official occupation and education classification systems — not proprietary tags.
India
- NCSNational Career Service — occupation codes for all roles
- UGCUniversity Grants Commission — 8-level education classification
- DPDPDigital Personal Data Protection Act — data residency compliant
- ISO 8601Date normalization — YYYY-MM format throughout
Europe
- ESCOEuropean Skills, Competences, Qualifications and Occupations
- ISCEDInternational Standard Classification of Education (UNESCO)
- GDPRGeneral Data Protection Regulation — EU data compliance
- ISO 8601Date normalization — consistent across all EU locales
Why Distil
Distil vs. manual CV screening
FeatureManual / SpreadsheetDistil
CV Processing SpeedHours per batch of 50 CVs< 2 seconds per CV
Structured OutputNone — recruiter's memory20+ fields, normalized JSON
Qualification MappingInconsistent per recruiterUGC/ISCED level — consistent
Indian DegreesOften missed or misspelled30+ Indian degrees recognized
Audit TrailNoneEvery action logged
Search & FilterNot possible from raw CVsAny structured field queryable
Human ReviewAll CVs reviewed manuallyOnly flagged fields surfaced
ScaleBreaks at 100+ CVs/dayScales to thousands
FAQ
Common questions
What is Distil?
Distil is an AI-powered CV standardization platform built by Data Strive Labs. It takes CV files (PDF, DOCX, HTML) and extracts 20+ structured fields — name, contact details, experience, education, skills, certifications, and salary expectation — then maps them to Indian (NCS, UGC) or European (ESCO, ISCED) standards. It scores confidence per field and routes weak extractions to human review.
What file formats does Distil support?
Distil currently supports PDF, DOCX (Word), and HTML files. PDFs are parsed natively — OCR fallback is available for scanned or image-heavy documents.
How does Distil handle Indian CVs?
Indian CVs are structured differently — the education section almost always appears at the bottom, after objective, skills, experience, and projects. Distil uses an 8000-character input window (not 3000 like most tools) to ensure education is never cut off. It recognizes 30+ Indian degrees and uses rule-based UGC level assignment so any degree — listed or not — gets the correct education level.
What standards does Distil apply?
For India: NCS (National Career Service) for occupation codes, UGC (University Grants Commission) levels 3-8 for education, and ISO 8601 for dates. For Europe: ESCO for occupations, ISCED for education, and GDPR-compliant data handling. The country is configured per organisation.
How accurate is the AI extraction?
Distil scores confidence per field. Common fields like name, email, and phone consistently score 95%+. Education and experience accuracy depends on CV quality. Fields below the confidence threshold are flagged for human review — you only manually check what the AI wasn't sure about.
Can I re-run the pipeline on an old CV?
Yes. The Reprocess button on any candidate detail page re-runs the full AI pipeline on the already-stored file. This is useful after Distil's extraction prompts are improved, or when you want to apply updated standardization rules to existing CVs. The original file is never deleted.
How is access to Distil controlled?
Distil is invite-only. The superadmin invites specific email addresses from the Team Access page. Invited users receive a branded email from noreply@datastrivelabs.in and can sign in via magic link or Google. Uninvited sign-in attempts are blocked with a clear message.
Who built Distil?
Distil is designed and built by Data Strive Labs, a custom AI and software company based in Hyderabad, India. It is live at distil.datastrivelabs.in and currently in an invite-only beta.
Ready to build something that works?
Tell us your biggest operational bottleneck. We’ll show you what’s possible — free, no strings.
Start a conversationOr call +91 96422 99939