lease-abstraction

name	lease-abstraction
title	Commercial Lease Abstraction
description	Extract a structured abstract from a commercial lease PDF — tenant/landlord identification, premises description + RSF/USF, lease term + commencement/expiration, renewal options (count, notice windows, fair-market rent reset), base rent schedule + escalations (fixed %, CPI-linked, step-up, percentage rent breakpoints), operating expense structure (NNN, modified gross, full service, base year, expense stops, gross-ups), CAM provisions + caps + audit rights, tenant improvement allowance, security deposit (cash/LOC/burn-down), use restrictions, co-tenancy + exclusivity, assignment/sublet rights, recapture, termination/default provisions, kick-outs, options to expand/contract, holdover, insurance/indemnification, estoppel + SNDA. Outputs a standardized abstract JSON + CSV + Markdown summary + critical-dates calendar (rent escalation alerts at 60/30 days, option exercise windows, expiration alerts). 80-90% reduction in processing time per industry benchmarks. TRIGGER on "lease abstract", "lease abstraction", "CRE lease", "commercial lease summary", "lease key terms", "rent roll setup", "lease review", "tenant rep prep", "Argus import", "lease database", "lease admin".
author	tinh2
author_url	https://github.com/tinh2/skills-hub-registry/tree/main/analysis/lease-abstraction
license	MIT
version	0.1.0
execution_mode	open
jurisdiction	general
practice	real-estate
language	en

Commercial Lease Abstraction

You convert a lease PDF (and amendments) into a structured abstract. Manual lease abstraction takes 4-12 hours per document depending on complexity; AI-assisted reduces that to 20-40 minutes of human review on top of a generated draft. The standardization matters as much as the speed — every abstracted lease should fit the same schema so a portfolio can be queried.

============================================================ === PRE-FLIGHT ===

• Lease document(s): original lease PDF + ALL amendments + side letters + work letters. Missing amendments = wrong financial picture.
• Lease type: office, retail, industrial, ground lease, residential (CA Civil Code §1947 differs), medical office, mixed-use. Drives which clauses to expect.
• Output schema target: VTS, Visual Lease, MRI, Yardi, ProLease, Excel template, or generic JSON. Each system has slightly different field names.
• Reporting purpose: lease admin (operational), tenant rep (negotiation prep), landlord acquisition DD, ASC 842/IFRS 16 accounting capture.

Recovery:

• Missing amendments: refuse to publish a final abstract; output draft + flag "amendments missing — DO NOT rely on for accounting."
• For complex retail leases (% rent + CAM gross-ups + exclusivity webs), surface complexity to a human reviewer; the abstract is a draft.

============================================================ === PHASE 1: DOCUMENT INTAKE + OCR ===

PDF preprocessing:

• If image-only PDF, OCR via pytesseract or AWS Textract (better for table extraction).
• Split lease into named sections (definitions, basic terms, rent, additional rent, options, default, indemnity).
• Extract embedded tables (rent schedule, CAM exhibit) as structured data, not flowed text.

Persist lease.text.json with section anchors so the abstract can cite source page+section per extracted field.

VALIDATION: Page count > 0, no all-white pages, text extraction confidence per section ≥ 80%.

============================================================ === PHASE 2: STANDARDIZED FIELD EXTRACTION ===

Extract into this schema (camelCase JSON):

{
  "parties":        
            
   
     
     
     
     
     
  
   
               
    ...