Skip to content
Recruitment technology · Platform

Parseit

Recruitment clients lodge a screening job, bulk-upload candidate resumes and state which fields they need extracted. The service operator works the queue from a separate back office and returns the resumes sorted into matching and not-matching sets, which the client downloads as a ZIP. Behind it sits an offline parsing pipeline that reads the documents, falls back to OCR when a PDF has no text layer, and classifies each candidate against a screening keyword list.
DeliveredPHPMySQLBootstrap
2
separate applications
4
user roles
19
screens
3
stage job workflow

Delivered on shared hosting in PHP, which is what the client's environment supported.

The problem

Screening a few hundred resumes by hand is the bottleneck in agency recruiting, and the documents arrive as a mix of PDFs, Word files and scans with no text layer at all.

What was built

The working parts

Each of these exists in the delivered system, not on a roadmap.

01

Client portal

Lodge a screening job with matching keywords and the fields to extract, track lodged, active and completed counts, and manage your own staff accounts.

02

Operator back office

A separate application with its own login for the service provider: client account management, every client's jobs, and the queue of work in progress.

03

Bulk upload and download

Multi-file upload for a whole batch, and selected resumes zipped on the fly for download.

04

Matching and not-matching buckets

Processed resumes are returned into two labelled sets per job, so the recruiter reads the shortlist rather than the batch.

05

Document parsing with OCR fallback

Word documents convert to PDF headlessly, text is extracted directly, and anything that yields no contact details is re-run through OCR rather than silently dropped.

06

Field extraction

Name, email, phone and date of birth pulled from the document, with the client choosing per job which of those they want.

07

Multi-tenant account structure

Client companies own their own staff users, and operator staff are separated into admin and super-admin roles.

Screens

Where people actually work.

  • 01Client dashboard with job counters
  • 02Lodge a job with keywords and extract fields
  • 03Job detail with upload and result buckets
  • 04Client staff management
  • 05Operator dashboard across all clients
  • 06Client account management
  • 07Per-client job list and status control
Stack
PHPMySQLBootstrapjQueryDataTablesPythonOCRmyPDFcPanel
Integrations
  • LibreOffice headless for document conversion
  • pdfminer for text extraction
  • OCRmyPDF for scanned documents

Want one built around your operation?

Discovery takes one to two weeks and produces a process map, wireframes and a fixed quote. Useful documents even if you take them somewhere else.

We reply to every enquiry within one business day.