Server definition
- Hash
- sha256:7471803b927d41664d2346e70c2876ed676780a581f34b1f93637f880e493383
- What it is
- What a remote MCP server returned when asked what it offers: 8 tools
The blob, as servednamed by its sha256
{
"instructions": "bharatlas MCP server (hosted, read-only). India's open geo data: curated layers (state to village boundaries, city wards, forests, hospitals, highways, rivers, canals, groundwater, agro-climatic zones, floods, seismic zones) plus community submissions.\n\nWorkflow patterns:\n\n- **Discovery**: start with list_layers or list_categories to find relevant layers. Use the q parameter for text search.\n- **Schema first**: always call get_layer_schema BEFORE query_layer. Column names vary per layer (e.g. \"state\" vs \"State_LGD\" vs \"stname\"). The schema shows exact names and sample values.\n- **Filtering**: query_layer where is an exact, case-insensitive match (no partial or substring match). Pass column=value pairs; check the schema for the right column name and value format. Columns the schema marks as type \"nested\" (structs, lists, maps, e.g. Overture names/addresses) can be selected but not used in where or group_by (the API returns 400 listing the flat columns).\n- **Counting**: use group_by to count features by any column. Example: group_by \"category\" on the gs_wildlife layer returns national parks vs sanctuaries.\n- **Totals**: add sum (numeric columns) to group_by to total values per group, e.g. group_by \"district\" + sum [\"area_ha\"] rolls block-level areas up to districts. Percentages are not additive: sum the underlying amounts and recompute shares.\n- **Location queries**: locate returns all admin boundaries + zones at a lat/lng, and (for the ~30 covered cities) auto-includes the municipal ward layer that contains the point — so \"what ward is this?\" just works with lat/lng alone. Use it to answer \"what state/district/ward is this point in?\"\n- **Spatial joins** (multi-step): to answer \"which X are in Y?\" when layers don't share a common column:\n 1. Use locate to find the admin context (state, district) of the target area\n 2. Use query_layer with a where filter to narrow the other layer by that admin context\n 3. For precise spatial containment, test individual points via locate against the target layer\n- **Community layers**: list_submissions returns user-contributed datasets. These work with query_layer and get_layer_schema exactly like curated layers.\n- **Downloads**: get_layer_detail returns direct URLs for the formats a layer has (parquet and pmtiles always; geojson, kml and shapefile for most).\n- **Large layers**: overture_places_india has 4.4M places. Always narrow it: where on basic_category (e.g. \"hospital\", \"restaurant\") or name, group_by, or nearby for places around a point. Whole-table filters and group_by take several seconds; nearby is the fast path for \"near me\".\n- **Honor the terms**: every layer and community submission carries its own source and open licence. When you use, quote, or publish results from a layer, credit its source and respect that licence; get_layer_detail returns the attribution and licence, and each /view page's Dataset JSON-LD carries a creditText to reproduce. A layer whose answer has \"non_commercial\": true is for non-commercial use only: say so whenever you use or recommend it, quote its \"licence_notice\", and never suggest it for a product, a client or paid work.\n\nSource preference (multiple layers often cover the same level):\n- **Admin boundaries** (state, district, subdistrict, block, village): prefer LGD (lgd_*) as the authoritative source. SOI, Bhuvan, and geoBoundaries are cross-reference alternates with slightly different counts/boundaries. Mention alternates if the user asks about discrepancies.\n- **City wards**: each city has its own layer (wards_chennai, wards_pune, etc.). Some cities have multiple vintages (e.g. Bengaluru has GBA 2025 + BBMP 2022 historical).\n- **Environment/infrastructure/health**: usually one source per layer. Check the source field.\n- **Community submissions**: ALWAYS check list_submissions alongside list_layers when searching. Community layers may cover areas or topics that curated layers don't. They work with query_layer and get_layer_schema identically.\n- When multiple sources exist, tell the user which one you're using and why. If results seem wrong, try the alternate source.\n\nCross-layer queries (combining data from different layers):\n- Questions often span multiple layers and categories. \"Which wards are near an international airport?\" needs the airports layer (transport) + a city ward layer (city-wards).\n- **Step 1**: identify all relevant layers using list_layers with different q/category filters.\n- **Step 2**: query one layer to get a geographic anchor (e.g. airport location → state/district).\n- **Step 3**: use that anchor to find the right layer in the other category (e.g. wards_bengaluru_gba for Karnataka airports).\n- **Step 4**: use locate or column filters to find overlapping features.\n- Never assume one layer is enough. Think about what layers need to be combined to answer the question fully.\n- Use locate as the bridge: find what's at a point across ALL relevant layers in one call by passing multiple layer IDs.\n- **Map user concepts to multiple layers.** Users say \"water bodies\" not \"wris_rivers.\" Translate:\n - \"water bodies/water\" → wris_rivers, wris_reservoirs, bp_wetlands, bp_ramsar, wris_basin, wris_subbasin, wris_canals\n - \"groundwater/aquifer/water table\" → cgwb_aquifers, cgwb_gw_extraction, mi6_wells_districts\n - \"wells/borewells/tubewells/irrigation pumps\" → mi6_wells_districts (census counts per district, 2017-18)\n - \"forests/green cover/ecology\" → soi_forests, gs_wildlife, bm_eco_zones, biogeographic_zones\n - \"agriculture/farming/cropping zones\" → agro_ecological_zones, agro_climatic_zones, corestack_lulc_blocks (land use and cropping intensity by block, modelled)\n - \"places/POIs/shops/restaurants/ATMs/schools/hospitals by name\" → overture_places_india (filter on name or basic_category; 4.4M rows, always add a where or group_by, or use nearby)\n - \"hazards/risks\" → seismic_zones, india_flood_inventory\n - \"health/medical\" → nic_health\n - \"boundaries/admin\" → lgd_states, lgd_districts, lgd_subdistricts, lgd_blocks, lgd_villages\n - \"roads/transport\" → gs_highways, airports\n When unsure, use list_layers with a broad q search and list_categories to discover what's available. Combine ALL relevant layers, not just the first match.\n- **Sky and air over a place**: once you have a coordinate (from locate, nearby or a layer row), the mugilu MCP (mugilu.live/mcp, tool conditions_at) gives the live air quality, heat, fog and warnings there. bharatlas answers where; mugilu answers what the sky is doing.\n\nColumn naming conventions (vary by source):\n- State: State_LGD (number), STNAME, stname, state, state_name\n- District: Dist_LGD (number), DTNAME, dtname, district, dtcode11\n- Block: Block_LGD, BNAME, block_name, blkcode11\n- Names are often UPPERCASE in LGD layers, mixed case in other sources\n\nAuthoring your own maps (collect.bharatlas.com: register a share link, moderate, import, publish) is not on this hosted server. It needs the npm package, which keeps share links on the user's machine: npx -y bharatlas-mcp.",
"tools": [
{
"description": "Get full metadata for a single layer: download URLs (parquet, pmtiles, geojson, kml, shapefile), attribution, license, feature count, and level metadata.",
"inputSchema": {
"properties": {
"layer_id": {
"description": "Layer ID",
"type": "string"
}
},
"required": [
"layer_id"
],
"type": "object"
},
"name": "get_layer_detail",
"outputSchema": null
},
{
"description": "Get column names, types, and sample values for a layer. Use BEFORE querying to discover what columns exist and what values they contain. Essential for knowing the right column names for filters.",
"inputSchema": {
"properties": {
"layer_id": {
"description": "Layer ID (e.g. lgd_states, airports, soi_forests, wards_chennai)",
"type": "string"
}
},
"required": [
"layer_id"
],
"type": "object"
},
"name": "get_layer_schema",
"outputSchema": null
},
{
"description": "List all data categories with layer counts.",
"inputSchema": {
"properties": {},
"type": "object"
},
"name": "list_categories",
"outputSchema": null
},
{
"description": "List available geo layers. Use to discover what data bharatlas has. Returns layer IDs, names, row counts, categories, and download formats. Filter by category (boundaries, city-wards, environment, water, agriculture, transport, etc.), admin level (state, district, block, village, etc.), source, or text search.",
"inputSchema": {
"properties": {
"category": {
"description": "Filter by category: boundaries, city-wards, people, environment, water, agriculture, transport, infrastructure, health-edu",
"type": "string"
},
"level": {
"description": "Filter by admin level: state, district, subdistrict, block, village, etc.",
"type": "string"
},
"limit": {
"description": "Max results (default 20)",
"type": "number"
},
"offset": {
"description": "Skip N results for pagination",
"type": "number"
},
"q": {
"description": "Text search across layer IDs, sources, and descriptions",
"type": "string"
},
"source": {
"description": "Filter by data source: LGD, SOI, Bhuvan, OpenCity, DataMeet, etc.",
"type": "string"
}
},
"type": "object"
},
"name": "list_layers",
"outputSchema": null
},
{
"description": "List community-submitted geo layers. These are user-contributed datasets under open licenses. Filter by category or search by name.",
"inputSchema": {
"properties": {
"category": {
"description": "Filter by category",
"type": "string"
},
"limit": {
"description": "Max results (default 20)",
"type": "number"
},
"offset": {
"description": "Pagination offset",
"type": "number"
},
"q": {
"description": "Search name and description",
"type": "string"
},
"sort": {
"description": "Sort order (default: recent)",
"enum": [
"recent",
"useful"
],
"type": "string"
}
},
"type": "object"
},
"name": "list_submissions",
"outputSchema": null
},
{
"description": "Given a latitude/longitude, find which administrative boundaries, zones, and regions contain that point. Returns state, district, subdistrict, block, parliament/assembly constituency, pincode, seismic zone, high court jurisdiction, the municipal ward (for the ~30 cities with ward data), and more. The 'where am I?' tool. Also accepts specific layer IDs to check.",
"inputSchema": {
"properties": {
"lat": {
"description": "Latitude (6 to 38, India bounding box)",
"type": "number"
},
"layers": {
"description": "Specific layer IDs to check. Omit to use the default: 13 essential admin/zone layers PLUS the city ward layer(s) covering the point (for the ~30 cities with ward data). Passing this replaces the default entirely (no ward auto-include).",
"items": {
"type": "string"
},
"type": "array"
},
"lng": {
"description": "Longitude (68 to 98, India bounding box)",
"type": "number"
}
},
"required": [
"lat",
"lng"
],
"type": "object"
},
"name": "locate",
"outputSchema": null
},
{
"description": "Find features from a layer that are near a given point, nearest first, within radius_km. Reads the layer's parquet bounding boxes and ranks features by distance to their centres; works for points, polygons and lines. Returns the true count within the radius plus the nearest `limit`. Example: 'reservoirs near Bengaluru' -> layer_id: 'wris_reservoirs', lat: 12.97, lng: 77.59, radius_km: 50.",
"inputSchema": {
"properties": {
"lat": {
"description": "Center point latitude",
"type": "number"
},
"layer_id": {
"description": "Layer to search for nearby features",
"type": "string"
},
"limit": {
"description": "Max results (default 20)",
"type": "number"
},
"lng": {
"description": "Center point longitude",
"type": "number"
},
"radius_km": {
"description": "Search radius in kilometres (default 25, max 200)",
"type": "number"
}
},
"required": [
"layer_id",
"lat",
"lng"
],
"type": "object"
},
"name": "nearby",
"outputSchema": null
},
{
"description": "Query a layer's data with filters and grouping. Reads the parquet file at runtime. Supports: select specific columns, filter by column values, group by a column to get counts, and sum numeric columns per group (e.g. roll block areas up to district totals). Use get_layer_schema first to discover column names. Examples: 'airports in Karnataka' -> where: {state: 'KA'}, 'forest types' -> group_by: 'type', 'health facilities in Bihar' -> where: {state: 'BIHAR'}, select: ['name', 'type'].",
"inputSchema": {
"properties": {
"group_by": {
"description": "Column to group by. Returns {value: count} instead of rows.",
"type": "string"
},
"include_centroid": {
"description": "Include _lat/_lng centroid coordinates (defaults to true). Set false to exclude.",
"type": "boolean"
},
"layer_id": {
"description": "Layer ID to query",
"type": "string"
},
"limit": {
"description": "Max rows (default 100, max 1000)",
"type": "number"
},
"select": {
"description": "Column names to return. Omit for all non-geometry columns.",
"items": {
"type": "string"
},
"type": "array"
},
"sum": {
"description": "Numeric columns to total per group (requires group_by). Adds {sums: {group: {col: total}}}. Use this to roll finer units up, e.g. block-level areas to district totals. Shares/percentages are not additive: sum the area columns and recompute shares.",
"items": {
"type": "string"
},
"type": "array"
},
"where": {
"additionalProperties": {
"type": "string"
},
"description": "Filter conditions as {column: value} pairs. Exact, case-insensitive match on flat columns (not partial). Values may contain commas.",
"type": "object"
}
},
"required": [
"layer_id"
],
"type": "object"
},
"name": "query_layer",
"outputSchema": null
}
]
}Verify it yourself
curl -s https://api.teppi.xyz/v1/evidence/sha256:7471803b927d41664d2346e70c2876ed676780a581f34b1f93637f880e493383 | sha256sum