Veraset MCP. Run SQL and Manage Geospatial Data Drops in Chat.

Look, you're dealing with billions of geolocation and mobility records from Veraset. You don't want to context switch between AWS consoles or some SQL client just to analyze foot traffic density or POI patterns; your AI agent handles that right here in the chat window.

When you connect your agent to this server, you gain direct access to massive datasets without leaving your workflow. It’s like treating a whole geospatial data warehouse—all those mobility and Points of Interest records—like an accessible database through simple commands.

Getting Started: Checking What You Got

You need to know what's in the bin before you start throwing code at it. Use list_mobility_datasets to get a full rundown of every dataset package Veraset has ready for querying. Once you pick one, you can run get_dataset_metadata to pull technical details about that whole mobility package. To really understand what's inside, use get_dataset_schema; this tells you the exact column names and data types so you know exactly how to write your SQL against it.

Before writing a single line of code, check out get_dataset_sample. This pulls the first few rows—a quick peek at the raw data format—so you can validate that everything looks right before you waste compute cycles. If you need more technical depth on a specific dataset, get_dataset_metadata gives you those specs.

Running and Managing Complex Queries

When you’re ready to calculate location aggregates or run complex analyses across the Veraset data, use execute_sql_query. This starts a new ANSI SQL job against the mobility dataset and immediately returns a unique job ID. Because these queries are massive—we're talking billions of records—you need to monitor them. Use get_query_status to check on the progress; it shows you what percentage is done and how many total bytes it’s scanned so far.

When that number hits 100%, use get_query_results to pull all the final rows from the job.

If a query runs too long or you realize you took a wrong turn, don't sweat it. You can immediately stop an executing SQL task using cancel_running_query. It kills the expensive process instantly so you don’t waste resources.

Handling Bulk Downloads: S3 Deliveries

Sometimes you need more than just a few rows—you need the whole damn thing for another system. The server manages bulk data drops stored in S3 buckets, which means it's all about secure downloads. First, run list_s3_delivery_folders to see every available S3 prefix where your scheduled, massive data packets are sitting.

Once you know the folder, use generate_download_link to create a temporary, secure URL. This link lets you download specific S3 files—usually Parquet bulk drops—without ever having to mess with AWS credentials or console navigation.