RERA and Project Document Data Extraction

• Head of Business Development or Land Acquisition
• Market Research Analyst at a developer or brokerage
• Construction Finance Credit Manager
• RERA Compliance Manager

State RERA portals hold the most reliable public record of what is being built: promoters, approvals, unit counts, bookings and declared completion dates. But each state publishes it differently, much of it sits in scanned PDFs, and analysts copy it into Excel by hand, one project at a time.

We build a collector for each state RERA portal you care about, plus a document pipeline for the files attached to each project. Browser automation walks the listings at a polite pace, captures project detail pages and downloads filed documents. OCR and a large language model extract the fields you define, such as promoter, registration validity, land area, towers, unit mix, booked units and declared completion date, and normalise them across states. Each run is compared with the previous one, so changes surface as a list rather than a surprise. Low-confidence values go to an analyst before the dataset is published to Excel, a database or your BI tool.

• Per-state collectors for RERA project listings and detail pages, scheduled weekly or monthly
• Download and classification of filed documents: registration certificate, approvals, architect and engineer certificates, quarterly progress reports
• Extraction of unit mix, carpet areas, booked and unsold units, and declared and revised completion dates
• Normalisation of field names, area units and promoter names across states, with de-duplication by registration number
• Quarter-on-quarter change log for extensions, new phases and inventory movement
• Source link and page reference stored with every extracted value

1. 01Agree scope and fieldsChoose states, districts and project types, and fix the field list and output format with your analysts.
2. 02Collect listings and documentsScheduled browser automation walks the portal listings within agreed rate limits and saves each project page and its attached documents.
3. 03Extract and normaliseOCR and an LLM read pages and PDFs, extract the agreed fields and map them to a common schema across states.
4. 04Check confidenceValues below threshold, or that conflict between the project page and a filed PDF, are queued for an analyst.
5. 05Compare with the last runA diff engine marks new projects, revised completion dates, added towers and changes in booked units.
6. 06Analyst reviewAn analyst confirms flagged values and notable changes, with the source page shown next to each field.
7. 07Publish the datasetThe checked dataset is exported to Excel, written to a database or refreshed in your dashboard.

1. Collection
   • Playwright collectors per state portal
   • Scheduler with rate limiting and retries
   • Raw page and PDF archive
2. Document understanding
   • OCR with table detection
   • Document type classifier
   • LLM field extractor with per-state schemas
3. Normalisation
   • State-to-common field mapping
   • Area unit and date normalisation
   • De-duplication by registration number
4. Review and change tracking
   • Run-to-run diff engine
   • Analyst review screen with source links
5. Delivery
   • Excel export
   • Database or API
   • BI dashboard refresh

Analysts confirm any value below the confidence threshold or in conflict between a project page and its filed documents, and review notable changes before a refreshed dataset is published. Where a portal requires a login or captcha, a person completes that step.

• State RERA portals (public pages only)
• Excel or Google Sheets
• PostgreSQL or your data warehouse
• Power BI, Tableau or Metabase
• Internal land acquisition or credit memo templates

• Collects only publicly disclosed pages, at a request rate agreed in advance and in line with each portal's terms
• No captcha or login bypass; those steps stay with a person
• Every value stored with its source URL and capture date
• Conflicting values are flagged, never silently overwritten
• Personal details that appear in filings are limited to what the analysis needs

• RERA disclosures are public, but personal data inside them is still handled with DPDP Act 2023 principles of purpose limitation and minimisation.
• Collection schedules and volumes are designed to avoid placing load on government portals.

• Real Estate & Construction

  Sale Agreement and Title Document Review

  Check each agreement for sale against the booking form, cost sheet and RERA record, and build a title chain from land documents with gaps flagged for your advocate.

• Real Estate & Construction

  Construction Progress Monitoring from Site Photos

  Turn geotagged site photos, drone images and time-lapse cameras into tower-by-tower progress status against the schedule, with photo evidence a site engineer confirms.

• Real Estate & Construction

  Customer Demand Letter and Receipt Reconciliation

  Match stage-linked demand letters with bank credits, home-loan disbursements, TDS certificates and GST, unit by unit, and give the CRM team an accurate statement of account.

• NBFCs & Lending

  Property and Title Document Summarisation for Home Loans

  Read sale deeds, encumbrance certificates, revenue records and approvals for home loans and LAP, build a chain-of-title summary and flag gaps for the advocate and credit team.