dank-mcp
by AgentDank
dank-mcp is a Model Context Protocol (MCP) server that responds to questions about supported Cannabis Datasets. It is designed for educational and legal purposes, allowing users to query cannabis-related data using tools like Claude Desktop.
Last updated: N/A
What is dank-mcp?
dank-mcp is an MCP server that provides access to cannabis datasets. It fetches, cleans, and stores data in a DuckDB database, then serves this data through a custom MCP server, allowing LLMs like Claude to query it.
How to use dank-mcp?
To use dank-mcp, you need to install it using go install github.com/AgentDank/dank-mcp@latest. Then, configure your LLM host program (e.g., Claude Desktop) to use the server by providing the path to the executable and any necessary arguments in a configuration file. Example configurations are provided in the README.
Key features of dank-mcp
Access to cannabis datasets
DuckDB integration
MCP server functionality
Data cleaning and preprocessing
Command-line interface
Use cases of dank-mcp
Finding specific cannabis products
Analyzing cannabis data
Integrating cannabis data with LLMs
Educational purposes
Legal compliance
FAQ from dank-mcp
What datasets are supported?
What datasets are supported?
Currently, the US CT Medical Marijuana and Adult Use Cannabis Brand Registry is supported.
How do I install dank-mcp?
How do I install dank-mcp?
The preferred method is using go install github.com/AgentDank/dank-mcp@latest.
How do I configure Claude Desktop to use dank-mcp?
How do I configure Claude Desktop to use dank-mcp?
Follow the Claude Desktop configuration tutorial and substitute the provided mcp-config.json with the correct path to your dank-mcp executable.
Where can I find snapshots of the datasets?
Where can I find snapshots of the datasets?
Snapshots are available in the AgentDank dank-data repository.
How is the data cleaned?
How is the data cleaned?
The data cleaning process involves removing weird characters, treating 'trace' amounts as 0, and removing rows with ridiculous data. The specific cleaning logic can be inspected in the code.