Open-source tools worth testing in real estate.

mvanhorn/last30days-skill

/last30days

Light technical setup · License: MIT

An AI-agent research skill that searches recent discussion across Reddit, X, YouTube, Hacker News, GitHub and other public sources, then produces a cited synthesis.

Where it can help CRE

Test tenant and consumer sentiment, investigate a vendor before a call, or find the live debate around a submarket before commissioning formal research.

Check before relying on it

Coverage depends on platform access, credentials and public discussion. Treat engagement as a lead to investigate, not as market evidence by itself.

Panniantong/Agent-Reach

Agent Reach

Light technical setup · License: MIT

A command-line setup layer that connects AI agents to web pages, social platforms, video transcripts, GitHub and RSS through a collection of upstream tools.

Where it can help CRE

Give a research agent more places to look when tracking a proptech company, public project, employer, tenant category or local issue.

Check before relying on it

It installs and coordinates several tools, and some channels use local browser cookies. Review every dependency, permission and platform term before using it.

firecrawl/firecrawl

Firecrawl

Developer setup · License: AGPL-3.0

A web data API and self-hostable service for searching, crawling and converting web pages into structured, model-ready content.

Where it can help CRE

Monitor public planning pages, economic-development sites, broker inventory pages or vendor documentation when the source permits automated access.

Check before relying on it

A crawler is not permission. Respect robots rules, site terms, privacy, copyright and data licenses; the hosted service can also carry usage costs.

docling-project/docling

Docling

Light technical setup · License: MIT

A local-capable document parser for PDF, Office files, images, email and other formats, with layout, table, formula and OCR support.

Where it can help CRE

Build a review pipeline for offering memoranda, leases, engineering reports and financial packages while retaining more document structure than plain text extraction.

Check before relying on it

Parsing is not validation. Reconcile material fields to page images and source documents, especially tables, footnotes and poor scans.

microsoft/markitdown

MarkItDown

Light technical setup · License: MIT

A lightweight Python utility that converts PDF, PowerPoint, Word, Excel, images, audio, HTML and structured text into Markdown for analysis workflows.

Where it can help CRE

Create a quick, readable input layer for an AI review of mixed deal-room files or a searchable internal knowledge base.

Check before relying on it

It is optimized for text analysis rather than high-fidelity reproduction, and it runs with the current process permissions. Use narrow access and sanitized inputs.

ocrmypdf/OCRmyPDF

OCRmyPDF

Light technical setup · License: MPL-2.0

A focused tool that adds a searchable OCR text layer to scanned PDF files while preserving the original page images.

Where it can help CRE

Make legacy leases, title documents, surveys, invoices and diligence reports searchable before extraction or review.

Check before relying on it

OCR can confuse digits, columns and low-quality handwriting. Keep the original image and visually verify every material term or number.

qgis/QGIS

QGIS

Desktop app · License: GPL-2.0-or-later

A full-featured desktop geographic information system for spatial data management, analysis, cartography and repeatable processing models.

Where it can help CRE

Combine parcels, zoning, flood, demographics, transit and portfolio data for site screening, market exhibits and acquisition diligence.

Check before relying on it

The map is only as current and legally usable as its layers. Preserve source dates, coordinate systems and licensing with every conclusion.

keplergl/kepler.gl

Kepler.gl

Light technical setup · License: MIT

A browser-based geospatial visualization tool designed to explore and present large location datasets.

Where it can help CRE

Turn a property, comp, tenant or mobility dataset into an interactive map for fast pattern-finding and stakeholder review.

Check before relying on it

A persuasive map can overstate a weak dataset. Define the denominator, time period and missing records before interpreting clusters.

gboeing/osmnx

OSMnx

Developer setup · License: MIT

A Python package for downloading, modeling, analyzing and visualizing street networks and other OpenStreetMap features.

Where it can help CRE

Study walkability, network distance, intersections and access patterns around a site or across a portfolio.

Check before relying on it

OpenStreetMap completeness varies by place, and network analysis does not replace a traffic, access or engineering study.

OvertureMaps/data

Overture Maps Data

Developer setup · License: MIT code; data licenses vary

An open map dataset organized around places, buildings, addresses, transportation and administrative boundaries.

Where it can help CRE

Prototype place, building and transportation analysis across markets without starting from separate local datasets in every geography.

Check before relying on it

Coverage, release dates and theme-specific licenses still need review. Match records back to authoritative local sources for material decisions.

ahacker-1/cre-agent-skills

CRE Agent Skills

Light technical setup · License: Apache-2.0

A community-built collection of AI-agent skills and reference files for CRE underwriting, diligence, financing, brokerage, legal and asset-management workflows.

Where it can help CRE

Study how a CRE workflow can be decomposed into repeatable agent instructions, then test selected pieces against your own models and standards.

Check before relying on it

This is a useful project to test, not an approved underwriting standard. Independently verify formulas, benchmarks, legal boundaries and every deal conclusion.

Anexen/pyxirr

PyXIRR

Developer setup · License: Unlicense

A fast financial-function library for XIRR, XNPV, IRR, NPV, day-count conventions and private-equity performance calculations.

Where it can help CRE

Recompute irregular property or fund cash flows in Python and create independent checks around spreadsheet return calculations.

Check before relying on it

A correct function can still receive the wrong dates, signs or cash flows. Test edge cases, including multiple-IRR situations, against the model of record.

stefanoamorelli/fred-mcp-server

FRED MCP Server

Developer setup · License: AGPL-3.0

An MCP server that lets compatible AI agents search and retrieve Federal Reserve Economic Data series.

Where it can help CRE

Bring rates, labor, inflation, construction and local economic series into a sourced market brief or underwriting context workflow.

Check before relying on it

Series definitions, vintages and transformations matter. Record the series ID and observation date, and do not let an agent silently choose the proxy.

duckdb/duckdb

DuckDB

Developer setup · License: MIT

An in-process analytical SQL database that can query local files and columnar data without standing up a database server.

Where it can help CRE

Join large rent, comp, lease, utility or portfolio files locally and make a repeatable analysis instead of a fragile spreadsheet chain.

Check before relying on it

Local execution reduces infrastructure, not governance. Define field meanings, access controls and reconciliation checks before scaling a workflow.

apache/superset

Apache Superset

Developer setup · License: Apache-2.0

A self-hostable data exploration and visualization platform for building charts, dashboards and governed analytical views.

Where it can help CRE

Publish portfolio, leasing, collections, capex or pipeline dashboards once the source data and business definitions are stable.

Check before relying on it

A dashboard is a distribution layer, not a source of truth. Secure the deployment and reconcile every KPI to the system of record.