Summary
Toronto's open-data catalog becomes searchable and queryable from the assistant, down to individual dataset records.
The 10 tools read the catalog: search and relevance tools find datasets, the record tools pull rows from their resources, and the analysis tools report structure and update cadence. Everything is read access to a public open-data service.
What it is
A server over Toronto's open-data catalog: its 10 tools search and list datasets, read records from a dataset's resources, and analyze dataset structure and update frequency.
What you get
- Search, list, and rank datasets by relevance (search_datasets, list_datasets, find_relevant_datasets)
- Read records from a package's datastore resources (get_package, get_first_datastore_resource_records, get_resource_records)
- Analyze a dataset's structure and how often it updates (analyze_dataset_structure, analyze_dataset_updates)
- List catalog categories and summarize matching datasets (get_data_categories, get_dataset_insights)
Requirements
Network access to Toronto's open-data service.
Setup effort
One command — npx mcp-remote https://toronto-mcp.s-a62.workers.dev/sse