DaedalMap is an open geographic query engine whose goal is to ask place-based questions in natural language and get answers on a map. It is described as geographic interoperability infrastructure rather than a single data visualization tool.
Purpose and positioning DaedalMap provides geographic interoperability capabilities: through the loc_id identity model, versioned boundary geometries, published cross-geographic-system mappings, and maintained data packs, it resolves coordinates and reference codes into administrative matches, pulls compatible shapes and data, and follows cross-mappings into other systems such as census tracts, postal areas, watersheds, and electoral districts. It covers disaster, population, economic, and climate data, aggregating more than 40 sources.
Main capabilities
- Geographic identity model: loc_id is DaedalMap's geographic identity model, and the administrative backbone is the primary hierarchy and default connection path. Other geographic families such as postal areas, watersheds, tribal areas, and marine areas retain their own loc_id identities. Published cross-mappings connect them to the backbone and to each other, and each row of a cross-mapping carries a relationship type, overlap weight, source, and version year.
- Geometry data: Provides a global baseline plus more than 8 countries in greater depth. The same geographic tools work globally down to Admin 2. Country versions add deeper administrative levels and reference families for Australia, Brazil, Canada, France, Germany, Mexico, the United Kingdom, the United States, and others.
- Data packs: More than 20 maintained data packs covering natural hazards (earthquakes, tsunamis, volcanoes, hurricanes, tornadoes, wildfires, floods, weather alerts), disaster risk and environmental burden, economic and business indicators, currency exchange rates, population, and climate.
- Access methods: Remote MCP server, HTTP API, map application, downloadable files, and this repository as the open runtime behind all of them.
How to use The hosted MCP server requires no installation or account for discovery. Any client that supports streamable HTTP MCP can use the same URL. It is recommended to start with get_tool_help, with topic='geometry' for geographic tasks; or use get_catalog, get_pack, and get_data in sequence to obtain data. Each tool accepts strict JSON parameters, the calling model converts user questions into these parameters, and the server returns typed errors with recovery guidance when a call is malformed.
Self-hosting This repository is the open runtime: FastAPI server, MCP and HTTP API, and map frontend. It can be used forself-hosting, running your own data, or extending the engine. Set the data folder in .env; if DATA_ROOT is left empty, the default application data folder is used. No cloud storage, database, or account is required. The repository does not include a data folder; docs/DATA_INSTALLATION.md explains how to install data packs. The application opens at http://localhost:7000 , and the same MCP server is located at http://localhost:7000/mcp。发现路由包括 GET /api/v1/guide, GET /api/v1/catalog, GET /api/v1/packs/{pack_id}, POST /api/v1/query/dataset. The model API key is optional; OPENAI_API_KEY or ANTHROPIC_API_KEY only supports the built-in chat panel; MCP clients bring their own model. Unless a commercial validator is configured, self-hosted instances will return commercial_access_unavailable on paid channels. In cloud data mode, the runtime locally caches small metadata files and queries Parquet in object storage via DuckDB httpfs.
Use cases Suitable for developers, researchers, and self-hosted users who need location-based question answering, geospatial data interoperability, administrative boundary resolution, cross-geosystem connectivity, and thematic data analysis in areas such as disasters, population, economy, and climate. It can be integrated into agent workflows via MCP, or used to build applications via the HTTP API.
Limitations and billing Discovery and small geographic lookups are free. Larger point/reference batches and some data packs are billed by usage: paid calls return the exact price before any charge, and payment is made via account credit with an API key. get_catalog reports the access channel for each data pack. The tool list is the complete public roster, not a roadmap; the bulk export/job builder and direct upstream live source wrappers remain internal or paused and are intentionally absent from tools/list. Data packs and geometry data carry their own source licenses, listed in each catalog entry.
Other information The project is MIT licensed. Official offerings include the website, app, agent documentation, geometry catalog, data packs, downloads page, file conversion tools, and llms.txt. Contact: support@daedalmap.com.
