fullstack / freelancer

SELECTED PROJECT / 01

Public records.
Useful prospects.

A landlord prospecting tool that turns scattered property data into a record you can actually work with.

County records find the owner. Grok researches a contact. The result lands in the database, with its source attached.

PythonDuckDBReactxAI / Grok
THE PATH TO A PROSPECTCO, USA
⌂
One resolved ownerCompany + property portfolio
✓
✳
Grok web researchFind a contact. Keep the source.
API
SOURCED CONTACT → DATABASE

A real workflow, open to inspect.One business record from the working project.

23linked properties
1focused research request
6/6persistence checks passed*

TRY IT / LIVE DATABASE

From owner to contact.

Research this business, inspect the evidence,
and download the saved result.

⌂ Rental Lead Manager / Grok edition Single-record demo

Contact research

✳ Powered by Grok

Search public business sources and save supported phone candidates to this record.

Loading the saved record…

Connecting to the demo database.

One live search per 24 hours. Repeat visits use the saved research.

Phone candidates are unverified. Review the cited source before outreach.

THE WORKFLOW

Each source has a job.

The larger project connects property data to a practical
prospecting workspace. Grok handles the research step.

01 / COLLECT

Start with the property.

County assessor feeds provide owners, mailing addresses, and parcels. Rental-listing scrapers add another source of contact information.

Public records + listing scrapers
02 / RESOLVE

Connect the owner.

Group properties into portfolios and match company names against Colorado’s business registry.

Owner matching + LLC resolution
03 / RESEARCH

Give Grok a focused job.

Search for the exact owner, check the business identity, and return phone candidates with source links.

xAI Responses API + web search
04 / USE

Keep the useful result.

Save new numbers without duplicates, retain the research, refresh the record, and export the contacts.

DuckDB + prospecting interface

UNDER THE HOOD

A small integration.
A useful result.

The original tool uses a Python API, DuckDB, and a React interface. The Grok fork adds a focused enrichment action to the existing record.

This public demo uses the same Grok request functions with its own database. The county and listing scrapers continue to fetch their source data directly.

Download the API script
grok_client.pyPython / xAI
# One request, grounded in web sources.
response = requests.post(
    "https://api.x.ai/v1/responses",
    headers=server_side_headers,
    json={
        "model": "grok-4.7",
        "input": owner_research_prompt,
        "tools": [{"type": "web_search"}],
        "max_tool_calls": 3,
        "store": False,
    },
)
# Save supported candidates, then refresh.
Abbreviated request · Download the full adapter above

Checked beyond the API response.

*The fork’s October 27, 2026 test confirmed immediate display, database writes, CSV export, page reload, disk persistence, and server restart. View the verification summary ↗