- SEARCH
- STRUCTURE
- DATA
- HUMAN
Short answer
How can I use AI to get ONS data?
Your AI searches the ONS catalog for specific topics, drills down into dataset dimensions, and pulls the exact numbers you need. You get raw statistical observations without manual browsing. Here is how the process works.
The results of your queries
Where your data lands
Depending on what you ask, your AI delivers different levels of data depth.
SEARCH
Targeted dataset discovery
The AI scans the 337+ available datasets to find the exact topic you're looking for. You get a list of relevant datasets instead of endless scrolling.
STRUCTURE
Dimension mapping
Your AI identifies how the data is organized, such as time periods or geographic regions. This ensures you know exactly what filters are available.
DATA
Precise observations
The AI pulls specific data points and values from the chosen dataset. You get the actual numbers you need for your analysis.
HUMAN
Final interpretation
The AI provides the raw data and metadata. You remain responsible for the final statistical analysis and reporting.
The workflow
What your AI does when you ask for data.
The AI moves from broad searches to specific data points using a sequence of specialized tools.
Locate datasets
The AI searches the ONS catalog to find datasets that match your specific keywords or topics.
search_datasetsInspect dataset structure
Once a dataset is found, the AI checks its metadata and available filters to understand the scope.
get_dataset_infoget_dimensionsIdentify filter options
The AI looks at specific dimensions to see what specific categories or timeframes are available for selection.
get_dimension_optionsExtract values
The AI pulls the actual statistical observations based on the dimensions you've selected.
get_observations
Try it
Copy these to start.
Use these prompts to begin your research.
Starting points
Replace the bracketed text with your specific research topics or timeframes.
UK ONS Connector
You're all set. Choose your MCP client and follow the setup instructions.
https://edge.vinkius.com/vk_preview_fvId9Opua7L9Nik9gCqSg8golmJupaDzpV9JVlBd/mcpClaude Desktop
Follow the steps below to connect in seconds.
- 1In Claude Desktop, open Settings → Connectors.
- 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
- 3Click Add and start a new chat — UK ONS capabilities are ready to use.
{
"mcpServers": {
"uk-ons-discovery-search-337-statistical-datasets-mcp": {
"url": "https://edge.vinkius.com/vk_preview_fvId9Opua7L9Nik9gCqSg8golmJupaDzpV9JVlBd/mcp"
}
}
}Search the ONS catalog for datasets related to [UK inflation rates].
Find the dimensions available in the [Consumer Price Index] dataset.
Get the statistical observations for [unemployment rates] in [London] for [2023].
Show me the available options for the [geography] dimension in the [population] dataset.
Claude
ChatGPT
Cursor
VS Code
Windsurf
Claude Code
JetBrains
Cline
Start here
Connect UK ONS Discovery once, then ask.
One click connects this MCP to your AI client. Your credentials stay encrypted, and you can start querying the ONS catalog immediately.
Connect UK ONS Discovery to your AIFAQ
How the ONS Discovery MCP behaves
- 01
Can the AI update or change ONS data?
No. The AI only reads data from the ONS catalog. It cannot modify any official statistics.
- 02
Can the AI download files directly to my computer?
No. The AI extracts observations and data points into your chat. You must manually save any data you want to keep.
- 03
How many datasets can the AI search?
The AI can search through all 337+ datasets available in the ONS catalog.
- 04
Does the AI perform the statistical modeling for me?
No. The AI pulls the raw observations and metadata. You are responsible for performing the actual statistical modeling or trend analysis.
- 05
Can I ask for data from specific years?
Yes, provided that the specific dataset contains those time dimensions.
More questions about UK ONS Discovery? The Connector page answers them. See everything the UK ONS Discovery Connector can do
Connect UK ONS to Claude, Cursor, ChatGPT & more
