The Grant PDF Parser is a tool for nonprofits and foundations to automatically extract useful budget and requirement details from grant-related documents (PDF, images, or text files). Grants fund the programs and people that drive impact, but the administrative work of parsing through long applications and agreements can be overwhelming. This project provides a simple, extensible way to turn unstructured grant documents into structured data for analysis, compliance, and project management.
-
Extracts structured data from PDF, image, and text-based grants.
-
OCR support with Tesseract
-
Uses LLM-powered parsing with LangChain.js and LangGraph.js
-
Outputs machine-readable JSON or CSV.
-
Captures key grant data:
- Grant name (multi-year support included)
- Start and end dates (broken down yearly)
- Project or program names
- Budget categories (salary, travel, supplies, fringe, indirect, etc.)
- Position-level allocations (position, program, budgeted amounts)
- Restrictions and compliance details (unrestricted vs. restricted, timesheet requirements, etc.)
-
Modular design: easy to extend for requirements, reporting metrics, or other grant-specific conditions.