Skip to content
Alex Reynolds

Tools & resources

Useful things, published in public.

Tools for navigating public information, guides for building better AI systems, and source code for anyone who wants to see how the machinery works.

5
MCP servers
7
Guides
82
Tools across them

Everything here began as a practical answer to a problem I kept encountering. The public versions are free to use, inspect, and adapt.

Run them yourself

Public records, in your own AI assistant.

Each of these is an MCP server. You install it once, and then you can ask your assistant about the data directly.

SC Elections & Campaign Finance

22 tools

GitHubnpm

South Carolina candidates, campaign finance, donors, certified results, and official disclosures.

npx -y sc-elections-mcp
Where the data comes from:
SC Ethics Commission (campaign finance, economic-interest statements), SC Votes / VREMS (ballot filings), and the certified results database.
Keep in mind:
Filed to run, on the ballot, and won are three different questions with three different sources. The server keeps them separate.

SC Legislature

18 tools

GitHubnpm

Bills, legislators, committees, calendars, and roughly 15,000 hearing and floor-session recordings.

npx -y sc-legislature-mcp
Where the data comes from:
Open States (bills, sponsors, legislators, current session), scstatehouse.gov (committees, delegations, meetings, calendars), and the Legislature's own video archive.
Keep in mind:
No other public service indexes the video archive, and direct MP4 links are available. The bill and legislator tools need a free Open States API key; everything else does not.

Greenville County Property

13 tools

GitHubnpm

Property, tax, deeds, zoning, historical court records, and GIS data for Greenville County, SC.

npx -y gc-property-search-mcp
Where the data comes from:
Greenville County public records: the assessor, the tax office, the Register of Deeds, and county GIS.
Keep in mind:
Deed search needs your own free Register of Deeds account. Everything else is open.

TN Property Records

7 tools

GitHubnpm

Statewide Tennessee assessment data and parcel geometry, covering 89 of 95 counties.

npx -y tn-property-search-mcp
Where the data comes from:
Tennessee county assessor data and parcel GIS services.
Keep in mind:
Six counties are not yet covered. No API key needed.

TN Elections & Campaign Finance

22 tools

GitHubnpm

Tennessee state-level campaign finance and official candidate filings, plus county campaign finance in 15 counties.

npx -y tn-elections-mcp
Where the data comes from:
Tennessee's online campaign finance system, Secretary of State candidate filings, and county election commission filings.
Keep in mind:
County coverage varies by what each commission publishes.

Installing one

The easiest path is Claude Desktop, which ships with everything it needs. Add the server under mcpServers in its config, using the command above. Claude Code takes a single line:

claude mcp add sc-elections -- npx -y sc-elections-mcp

Cursor, VS Code, and LM Studio all work the same way. One catch: VS Code calls the config key servers, where everyone else calls it mcpServers.

Claude on the web and on mobile only accepts servers running on a public URL, and these run on your own machine, so use the desktop app or Claude Code. ChatGPT can reach a server on your machine only through OpenAI's Secure MCP Tunnel, which is a developer setup.

Or let your assistant walk you through it

Paste this into Claude and answer its questions. It knows your setup better than a page of instructions can.

I want to use the SC Elections MCP server with you.

Package: sc-elections-mcp
Command: npx -y sc-elections-mcp
Requires: Node.js 20+
Docs: github.com/asreynolds1000/sc-elections-mcp
It provides 22 tools for South Carolina campaign
finance and election data.

Walk me through installing this in my setup, one
step at a time. First ask me which AI app I am
using and what computer I am on.

Swap the package name and description for whichever server you want.

In the browser

Audacy Downloader

Turn an Audacy episode page into a direct downloadable podcast audio file. Nothing to install.

Open tool →

Read a guide

Patterns learned through use, failure, and another pass.

Featured field guide

Help it help you

Three habits that move an AI chat past the search box: give it access, write it down, and ask for a blind review. Plus a step-by-step page for setting it up yourself.

Read the full guide →
01Give it access
02Write it down
03Ask for a blind review

Writing

AI writing: avoid the slop

Banned words, structural tells, and formatting habits that make machine-written text obvious.

Read guide →

Operations

Checkpointing for AI agents

How to stop losing useful work when a long agent session reaches its context limit.

Read guide →

Memory

Should you remember this?

Six questions to answer before adding another permanent rule or memory.

Read guide →

Architecture

A memory architecture for AI agents

A layered system for durable rules, lean indexes, project context, and on-demand knowledge.

Read guide →

Security

Don't let an LLM gate its own dangerous actions

A capability-level pattern for keeping irreversible actions outside the model's control.

Read guide →

Review

Ask for a blind review

How to check AI work before it goes out: a fresh reviewer that sees the work and the sources, but not your conversation.

Read guide →

Build from source

The working parts are open.

The servers above are published on GitHub and npm. Alongside them are a standalone project and smaller implementation patterns from client and research work.

Standalone project

API Cost Dashboard

A self-hosted SQLite dashboard for tracking actual billed usage across OpenAI, xAI, Gemini, and OpenRouter.

View on GitHub ↗

Patterns & field notes

Smaller lessons, written down.