# AgentNative Government and Public Data (full) AgentNative provides normalized official government datasets with provenance, deterministic aggregations, and a free paging plus aggregate preview. Paid access is a $5 pass (250 queries, 30 days) or $5/month (2,000 queries per billing period). ## Free paging cookbook 1. List query-ready data: GET https://agentnative.dennery.ai/v1/warehouse 2. Sample one dataset: GET https://agentnative.dennery.ai/v1/warehouse/sample/nyc-open-data-f9bf-2cp4 3. Page free rows: GET https://agentnative.dennery.ai/v1/warehouse/query/nyc-open-data-f9bf-2cp4?$top=25&$skip=0, then follow page.next_url while page.has_more is true (free ceiling: top 25, skip 475). 4. Consolidate without downloading: GET https://agentnative.dennery.ai/v1/warehouse/aggregate/nyc-open-data-f9bf-2cp4?group_by=school_name&metric=count&top=10 (free ceiling: top 10). 5. Check breadth first: GET https://agentnative.dennery.ai/v1/coverage and https://agentnative.dennery.ai/v1/sources. ## Court metadata cookbook SDNY publishes filing-feed metadata; CCJ publishes judgment links and citation metadata. Judgment document contents and PACER records are not imported. Verify provenance.source_url on every response. GET https://agentnative.dennery.ai/v1/coverage includes court_sources with per-court discovery status and published dataset counts; court_sources_limit is 100. 1. Discover court sources: GET https://agentnative.dennery.ai/v1/sources, select level court, then GET https://agentnative.dennery.ai/v1/datasets/discovered?source=ccj-judgments or source=nysd-ecf. 2. Sample three judgment metadata rows: GET https://agentnative.dennery.ai/v1/warehouse/sample/ccj-judgments-recent-e4f72b1f02825a793bdae5c1. 3. Page judgment metadata: GET https://agentnative.dennery.ai/v1/warehouse/query/ccj-judgments-recent-e4f72b1f02825a793bdae5c1?$top=2&$skip=0, then follow the exact page.next_url. 4. Group judgments by country: GET https://agentnative.dennery.ai/v1/warehouse/aggregate/ccj-judgments-recent-e4f72b1f02825a793bdae5c1?group_by=country&metric=count&top=10. Missing country values remain null. 5. For filings use dataset nysd-ecf-recent-filings-16f01255391820782fd11159 with the same sample/query routes; group_by=pubDate groups the published filing timestamps. Free bounds are top 25 and skip 475 for rows, top 10 for aggregate groups. Exact filters require paid bearer access. Paid bounds are top 100 and skip 5000 for rows and top 50 for aggregate groups. Use the shared query_imported_dataset and aggregate_imported_dataset MCP tools with these dataset IDs. No court-specific credentials are required for these public metadata datasets. ## Endpoints - GET https://agentnative.dennery.ai/health - GET https://agentnative.dennery.ai/ready - GET https://agentnative.dennery.ai/v1/catalog - GET https://agentnative.dennery.ai/v1/datasets/search - GET https://agentnative.dennery.ai/v1/sample/us-federal-register - POST https://agentnative.dennery.ai/v1/capabilities/request - POST https://agentnative.dennery.ai/v1/materializations - GET https://agentnative.dennery.ai/v1/materializations/{job} - POST https://agentnative.dennery.ai/v1/access/start - POST https://agentnative.dennery.ai/v1/access/claim/{claim} - GET https://agentnative.dennery.ai/v1/query/us-federal-register - GET https://agentnative.dennery.ai/v1/warehouse - GET https://agentnative.dennery.ai/v1/sources - GET https://agentnative.dennery.ai/v1/coverage - GET https://agentnative.dennery.ai/v1/datasets/discovered - GET https://agentnative.dennery.ai/v1/warehouse/sample/{dataset} - GET https://agentnative.dennery.ai/v1/warehouse/query/{dataset} - GET https://agentnative.dennery.ai/v1/warehouse/aggregate/{dataset} ## Protocols - MCP: https://agentnative.dennery.ai/mcp/government-data (tools: search_discovered_datasets, list_imported_datasets, sample_imported_dataset, query_imported_dataset, aggregate_imported_dataset, sample_us_federal_register, request_dataset_materialization, get_materialization_status) - A2A: https://agentnative.dennery.ai/.well-known/agent-card.json - OpenAPI: https://agentnative.dennery.ai/openapi.json