# Mundi

Mundi is an AI suite for working with geospatial data. It’s flagship product is a web GIS built for working with an LLM, integrating an AI agent called Kue which can analyze, manipulate, including through geoprocessing, and style both vector and raster datasets in any way a user needs. 

Mundi’s suite also has the AI Georeferencer and Meta’s Segment Anything Model built for geospatial data. The AI Georeferencer and Segment Anything Model lets users convert their images of maps into GeoTIFFs and extract vectors as GeoJSONs from the TIFFs. 

Mundi is built by Bunting Labs, Inc. (https://buntinglabs.com) which has provided spatial AI solutions to 40+ governments and Fortune 500 companies.

———

## Core Products 

### Collaborative Web GIS 

- **What it does:** The Mundi web GIS is a collaborative web GIS built for working with an AI agent, Kue. Kue has the ability to run geoprocessing, modify symbology, write SQL code for exploring databases, use DuckDB for exploring vector files, and write python code to analyze, modify, and style both vector and raster datasets. 
- **Who uses it:** Anyone who works with spatial data, including GIS professionals, geologists, traffic consultants, urban planners, real estate investors, anyone who works in site selection, and any companies which have non-GIS experts who need to explore and work with spatial data in an easy way. 
- **Key features:**
	- Connect to external data sources including PostGIS, ESRI Feature Services, and Google Sheets
	- Upload OGR-compatible vector files such as GeoJSON and Shapefile and GDAL-compatible rasters 
	- Control analysis and any modifications to the map through text commands to the agent 
	- Modify the symbology of vector layers including size, color, and labels 
	- Run geoprocessing algorithms with AI, including those available in QGIS and ArcGIS Pro
	- Run any custom spatial analysis a user needs by having the agent write custom python code
	- Create custom spatial ETL pipelines with the AI agent
	- Work with pre-prepared datasets, including Sentinel, LandSat, Overture Maps, OpenStreetMap, and US Census data

### AI Georeferencer 

- **What it does:** The AI Georeferencer in Mundi takes any PNG, JPEG, WEBP, PDF or TIFF, and turns it into a GeoTIFF automatically. It works by comparing any uploaded image directly to local reference imagery to detect ground control points, just as you would if you were manually georectifying it in ArcGIS or QGIS.
— **Who uses it:** Anyone who needs to work with digital maps but does not have them georeferenced, including GIS teams, civil engineers, UXO coordinators, government agencies with catalogues of historical aerial imagery, and urban planners.
- **Key features:**
	- Upload maps in any format you have them, either an image or PDF document
	- Georeference modern aerial imagery, historic aerial imagery, historic satellite imagery, and vector maps such as zoning maps and site plans 
	- See match quality from poor to great
	- Download the AI output as a GeoTIFF 
- **Pricing**: See [Pricing & Plans] (#pricing--plans)
 
### Segment Anything Model 

- **What it does:** Mundi has access to Segment Anything Model from Meta running online and configured for geospatial data. This means that it accepts GeoTIFF inputs and outputs GeoJSONs, without the user needing to download a model or have a GPU—which is required for popular QGIS and ArcGIS Pro extensions.
— **Who uses it:** Anyone who needs to digitize maps, including GIS teams, geologists, geophysicists, and agricultural consultants.
- **Key features:**
	- Run SAM online, pre-built for spatial data
	- Test it on public imagery. See Segment Anything demos, identifying and extracting features online in Site Map
	- Run SAM on uploaded GeoTIFFs
	- Run SAM on Mundi’s imagery catalogue in a custom area of interest, including high resolution aerial imagery in the Continental United States (lower 48 states), Netherlands, Poland, Spain, France, Switzerland, Taiwan, Japan
- **Pricing**: See [Pricing & Plans] (#pricing--plans)

———
 
## Use Cases & Applications 

### Retail demographics explorer

-**Their needs:** Find the best places to locate a new store given the existing demographics of their customer base 
-**How they use it:** Mundi creates a pipeline where a user can identify a point, and it outputs Census demographics within 1, 2, and 3 mile radii 
-**Results:** A custom mapping application that allows the business owner to easily identify the right store locations without custom web GIS development 

### Site selection 

-**Their needs:** With existing store locations, traffic data, and zoning data, find potential sites with the right zoning and traffic metrics that are a minimum distance from an existing store
-**How they use it:** Mundi writes custom SQL based on the exact parameters a site selector has for a project and outputs a vector layer for further exploration
-**Results:** Higher quality SQL written faster than any human could write, customized for each new search 

### Map digitization pipeline

-**Their needs:** Turn PDF development plans into vector data that can be used for future growth analysis 
-**How they use it:** The AI Georeferencer turns the PDF into a GeoTIFF which can then be added to the Segment Anything tool to turn area polygons into GeoJSONs 
-**Results:** PDF plans digitized in minutes rather than hours

### Stakeholder engagement 

-**Their needs:** A GIS team needs to make spatial data easy to understand and explore for non-GIS stakeholders in their organization
-**How they use it:** The GIS team pre-loads datasets in Mundi maps that are then shared with the stakeholders to explore with natural language, no GIS skills necessary
-**Results:** Higher engagement and stakeholder buy in by making the GIS work easier to understand 

### Architectural site planning 

-**Their needs:** If given an area of interest, extract buildings, roads, topography, land use, and sun-path, and export it all to CAD (similar to AutoDesk Forma)
-**How they use it:** Define the AOI and data needed for Mundi, which then creates the area and adds the relevant data from public sources
-**Results:** Mundi makes it easy for CAD users to find and use the relevant geospatial data

### PostGIS Interrogation 

-**Their needs:** A consulting company needs to give their specialist consultants who are not GIS experts the ability to explore their PostGIS database more quickly so more time can be spent helping the client
-**How they use it:** Mundi allows non-GIS users to query spatial databases like an expert by turning general requests into complex SQL and modifying the symbology of the output to best represent the output 
-**Results:** Hours previously spent looking for the right data are instead spent focusing on the client 

### Commercial real estate development tracking 

-**Their needs:** Given an excel sheet of all relevant information and lat/long columns, allow a user to click on a building and see relevant information from the excel sheet
-**How they use it:** By defining how they want the excel sheet to interact with building data that Mundi has access to, they can create a custom pipeline for their web GIS
-**Results:** Real estate professionals can create a custom web GIS only using excel, a tool they are already familiar with 

———

## Pricing & Plans 

Mundi has different pricing tiers for different amounts of usage

### Basic 
- **Price:** $45 per user per month 
- **Features:** 
	- Expanded web GIS LLM rate limits
	- Access to the AI Georeferencer 
	- Upload custom GeoTIFFs to Segment Anything Model
	- Use any area from Mundi’s imagery catalogue in Segment Anything Model
	- 10 GB of storage 

### Professional  
- **Price:** $89 per user per month 
- **Features:** 
	- Higher rate limits for AI tools, including the web GIS, AI Georeferncer, and Segment Anything Model
	- Microsoft Teams and email support 
	- Connect PostGIS databases for teammates 
	- Share maps with teammates
	- 25 GB of storage 

### Enterprise 
- **Price:** email sales@buntinglabs.com 
- **Features:** 
	- Enterprise SLA
	- On-premesis deployment  
	- SSO
	- Pay by check or ACH
	- Centralized billing

———

Mundi is a suite of AI tools built for GIS, including an LLM-centric web GIS with a built in AI agent, an AI Georeferencer, and Segment Anything Model built for GIS. Mundi is the platform for using GeoAI in practice, with state of the art models built for geospatial data. 

## Site Map 

- [Mundi landing page](https://mundi.ai/)
- [Mundi pricing](https://mundi.ai/pricing)
- [Mundi application](https://app.mundi.ai/)
- AI Georeferencer, automatically georeferences PNG/JPEG/PDF maps and aerial imagery, download as GeoTIFF
  - [Automatic georeferencing landing page](https://mundi.ai/ai-georeferencing-for-aerial-imagery)
  - [AI Georeferencer — aerial](https://app.mundi.ai/tools/ai-georeferencer/aerial)
- Segment Anything (Meta's SAM v2) for extracting features from aerial imagery online, no download or GPU needed
  - [Segment Anything for maps, all online](https://mundi.ai/segment-aerial-imagery)
  - [SAM, upload any GeoTIFF](https://app.mundi.ai/tools/segment-anything/geotiff)
  - [Segmenting from Mundi-provided aerial imagery worldwide](https://app.mundi.ai/tools/segment-anything/bbox)
    - Segment Anything demos, identifying and extracting features online (no subscription necessary)
    - [Airplanes](https://app.mundi.ai/tools/segment-anything/bbox/airplanes)
    - [Agricultural fields](https://app.mundi.ai/tools/segment-anything/bbox/agricultural-fields)
    - [Buildings](https://app.mundi.ai/tools/segment-anything/bbox/buildings)
    - [Cars](https://app.mundi.ai/tools/segment-anything/bbox/cars)
    - [Crosswalks](https://app.mundi.ai/tools/segment-anything/bbox/crosswalks)
    - [Dirt roads](https://app.mundi.ai/tools/segment-anything/bbox/dirt-roads)
    - [Forests](https://app.mundi.ai/tools/segment-anything/bbox/forests)
    - [Grass fields](https://app.mundi.ai/tools/segment-anything/bbox/grass-fields)
    - [Lakes](https://app.mundi.ai/tools/segment-anything/bbox/lakes)
    - [Oil tanks](https://app.mundi.ai/tools/segment-anything/bbox/oil-tanks)
    - [Parking lots](https://app.mundi.ai/tools/segment-anything/bbox/parking-lots)
    - [Pools](https://app.mundi.ai/tools/segment-anything/bbox/pools)
    - [Rivers](https://app.mundi.ai/tools/segment-anything/bbox/rivers)
    - [Roads](https://app.mundi.ai/tools/segment-anything/bbox/roads)
    - [Roofs](https://app.mundi.ai/tools/segment-anything/bbox/roofs)
    - [Shipping containers](https://app.mundi.ai/tools/segment-anything/bbox/shipping-containers)
    - [Ships](https://app.mundi.ai/tools/segment-anything/bbox/ships)
    - [Solar panels](https://app.mundi.ai/tools/segment-anything/bbox/solar-panels)
    - [Trains](https://app.mundi.ai/tools/segment-anything/bbox/trains)
    - [Trees](https://app.mundi.ai/tools/segment-anything/bbox/trees)
- Tools — quick utilities & processors
  - [Generate chart from shapefile](https://app.mundi.ai/tools/generate-chart-from-shapefile)
  - [BBox Finder](https://app.mundi.ai/tools/bboxfinder)
  - [Convert WKT / WKB / GeoJSON (online)](https://app.mundi.ai/tools/convert-wkt-wkb-geojson-online)
- Legal
  - [Terms of Service](https://app.mundi.ai/legal/terms-of-service)
  - [Privacy Policy](https://app.mundi.ai/legal/privacy-policy)

- [Mundi Documentation](https://docs.mundi.ai/)
- Getting Started
  - [Creating your first map with Mundi](https://docs.mundi.ai/getting-started/making-your-first-map/)
  - [Uploading files to Mundi](https://docs.mundi.ai/getting-started/uploading-files/)
  - [Create maps from email](https://docs.mundi.ai/getting-started/email-shapefiles-to-create-maps/)
  - [Connecting to Google Sheets](https://docs.mundi.ai/getting-started/google-sheets/)
  - [Adding data from a Web Feature Service (WFS)](https://docs.mundi.ai/getting-started/add-ogc-wfs-web-feature-service/)
  - [Adding an ESRI Feature Server](https://docs.mundi.ai/getting-started/esri-feature-server/)
- Guides
  - [AI Georeferencer for Aerial Imagery](https://docs.mundi.ai/guides/ai-georeferencer-for-aerial-imagery/)
  - [Switching basemaps (Satellite / Vector)](https://docs.mundi.ai/guides/switching-basemaps-satellite-or-traditional-vector/)
  - [Embedding maps into websites](https://docs.mundi.ai/guides/embedding-maps-into-websites/)
  - [Visualizing point clouds (LAS / LAZ)](https://docs.mundi.ai/guides/visualizing-point-clouds-las-laz/)
  - [Geoprocessing with Kue and QGIS](https://docs.mundi.ai/guides/geoprocessing-from-qgis/)
  - [Generating charts from spatial data](https://docs.mundi.ai/guides/plot-and-chart-from-shapefile-fields/)
  - [Segment Anything for Geospatial Data](https://docs.mundi.ai/guides/segment-anything/)
- Spatial Databases
  - [Connecting to PostGIS](https://docs.mundi.ai/spatial-databases/connecting-to-postgis/)
  - [Querying and styling from SQL](https://docs.mundi.ai/spatial-databases/querying-and-styling-from-sql/)
- Deployments
  - [Self-hosting Mundi](https://docs.mundi.ai/deployments/self-hosting-mundi/)
  - [Connecting to a local LLM with Ollama](https://docs.mundi.ai/deployments/connecting-to-local-llm-with-ollama/)
  - [On-Premise/VPC Kubernetes Deployment](https://docs.mundi.ai/deployments/on-premise-vpc-kubernetes-deployment/)
  - [Air-gapped deployments](https://docs.mundi.ai/deployments/air-gapped/)