io.github.cyanheads/socrata-mcp-server

Run SELECT-only SQL against a DataCanvas table populated by socrata_query_dataset. Columns SODA types as number (including aggregate aliases like count(*) as n) are staged as DOUBLE, so numeric comparisons work without a cast (year > 2020, amount < 500). Text and timestamp columns stay VARCHAR — compare times with CAST(date AS TIMESTAMP). Only works when CANVAS_PROVIDER_TYPE=duckdb is set. Use socrata_dataframe_describe to see registered tables and their schemas.

List known Socrata-powered government open-data portals with their domain, organization name, and approximate dataset count. The catalog is a curated list of 39 well-known portals; dataset counts are fetched from the Discovery API and cached for ~24 hours. Filtering is client-side substring match on the query parameter. Use this first when you do not know which portal to target, then pass the domain to socrata_find_datasets.

Fetch full metadata and column schema for a Socrata dataset by ID. Returns field names, data types, descriptions, row count, and licensing. Always call this before writing a socrata_query_dataset — the column types determine correct WHERE clause syntax: Number columns accept bare literals (year=2023) while Text columns require single-quoted strings (year='2023').

Execute a SoQL query against any dataset on any Socrata portal. Use the search parameter for quick full-text lookup, or combine select/where/group/having/order for full analytical control. Returns rows plus the assembled SoQL string so you can learn the pattern. Columns are referenced by API field name (field_name from socrata_get_dataset, e.g. cuisine_description), never the display label. All SODA 2.1 row values are strings even for numeric columns — check data_type from socrata_get_dataset to determine correct WHERE quoting: Number columns use bare literals (year=2023), Text columns use single-quoted strings (year=\'2023\'). To enumerate distinct values, use select="col, count(*) as n" with group="col" and order="n DESC". When CANVAS_PROVIDER_TYPE=duckdb and rows fill limit, up to 50,000 matching rows spill to a DataCanvas table whatever the limit: list its columns with socrata_dataframe_describe, then run SQL with socrata_dataframe_query.

List registered tables in a DataCanvas session — schema, row count, and column names. Shows what datasets are available for SQL queries via socrata_dataframe_query. Only meaningful when CANVAS_PROVIDER_TYPE=duckdb is set. Use after socrata_query_dataset spills a large result set to canvas.

Search for datasets across all Socrata-powered government open-data portals, or scope to one portal with the domain parameter. Returns dataset IDs, names, domains, update timestamps, and column_names — the API field names SoQL takes, not display labels. Use socrata_get_dataset to fetch the typed column schema before writing queries — column_names carry no type information.