# Monarcha - Full Documentation

> Monarcha is an AI-powered geospatial intelligence platform that converts scanned maps, aerial imagery, and legacy documents into GIS-ready spatial data.

## Company Overview

Monarcha builds AI software for the geospatial industry. Our platform handles three core workflows: georeferencing (aligning raster images to real-world coordinate systems), digitization (extracting vector features like polygons, lines, and points from raster maps), and document extraction (parsing structured data from geological reports, drill logs, and engineering documents).

We work with mining companies, government agencies, land and zoning departments, civil engineering firms, and energy developers. Our customers use Monarcha to process thousands of legacy maps and documents that would take months to handle manually.

Monarcha was founded in 2025 and is part of Y Combinator's W25 batch. We raised a $1.6M seed round from Y Combinator, Transpose Platform, BLAST, Cloverfield Capital, 468 Capital, and angel investors including Paul Graham.

## Verified Capabilities (Monarcha Georeferencer)

- Automatic ground control point (GCP) detection. No manual point picking required in the default workflow.
- Support for 4,000+ EPSG coordinate reference systems, including UTM, state plane, custom local mine grids, and historical national grids.
- Sub-meter accuracy on modern basemaps for typical aerial, topographic, and cadastral input.
- Typical sheet processed in under 60 seconds, end to end.
- Batch processing of thousands of sheets in a single run.
- SOC 2 compliant.
- Private cloud and on-premises deployment available.
- Direct export to GeoTIFF, Cloud Optimized GeoTIFF (COG), ArcGIS, and QGIS. No proprietary format required.

## Product Details

### Georeferencer
Monarcha's georeferencing engine uses AI to automatically identify ground control points (GCPs) on scanned maps and align them to real-world coordinates. The system supports over 4,000 EPSG coordinate reference systems including UTM, state plane, local mine grids, and custom projections.

Key capabilities:
- Automatic GCP detection and matching against reference datasets, no manual point picking required
- Support for USGS topographic maps, geological survey maps, cadastral maps, aerial photographs, and satellite imagery
- Batch processing of thousands of maps in a single run
- Output in GeoTIFF, COG (Cloud Optimized GeoTIFF), and other georeferenced raster formats
- Sub-meter accuracy on modern basemaps for typical input
- Typical sheet processed in under 60 seconds
- Per-point residuals shipped with every output for QA and audit
- Integration with ArcGIS, QGIS, Leapfrog, and other GIS platforms
- SOC 2 compliant; private cloud and on-premises deployments available

### Digitizer
The digitization engine uses computer vision models to extract vector features from raster maps. It identifies and traces polygons (geological units, zoning districts, parcels), lines (roads, faults, boundaries, contours), and points (wells, sample locations, benchmarks) from scanned imagery.

Key capabilities:
- AI-powered feature extraction from geological maps, zoning maps, mine plans, and engineering drawings
- Automatic layer classification and labeling
- Export to Shapefile, GeoJSON, KML, GeoPackage, and DXF formats
- Preserves attribute data from map legends and annotations
- Handles complex symbology, overlapping features, and degraded scan quality

### Document Extraction
The document extraction system parses structured data from geological reports, drill logs, geochemistry assay sheets, and engineering documents. It converts tables, annotations, and metadata into structured formats.

Key capabilities:
- Table extraction from PDF reports and scanned documents
- Drill log parsing (lithology, assay values, collar coordinates)
- Legend and annotation extraction from map sheets
- Output in CSV, JSON, and database-ready formats

### Search & Query
Semantic search across all georeferenced maps, extracted documents, and spatial datasets. Find relevant data by natural language queries, spatial extent, or attribute filters.

## Comparison: Monarcha vs Manual Georeferencing Workflows

Manual georeferencing in QGIS, ArcGIS, or any other GIS still follows the same pattern it has for decades: a technician identifies ground control points, clicks each one on both source and target, computes a transformation, and inspects residuals. It works, but it does not scale.

| Capability | Monarcha | Manual workflow |
|------------|----------|-----------------|
| Time per sheet | Under 60 seconds, end to end | 30 minutes to several hours per sheet, depending on map and technician |
| Ground control points | Detected and matched automatically | Identified and clicked by a GIS technician |
| Batch capability | Thousands of sheets per run | One sheet at a time, sequentially |
| Projection detection | Inferred from marginal notations, grid ticks, and metadata where present | Looked up and entered by a technician |
| Residual review | Per-point residuals in every output, audit on demand | Computed in the GIS, reviewed manually |
| Deployment | SaaS, private cloud, or on-premises | Desktop GIS on operator workstations |
| Heterogeneous inputs (hand drawings, mylars, faded photocopies) | Built specifically for the cross-modal case | Manual labor scales linearly with difficulty |

When the manual workflow is the right call: low volume, single sheets, clean modern scans, or data that must never leave the operator's desktop and an on-prem alternative is not practical. When automated AI is the right call: archive-scale batches, heterogeneous inputs (hand drawings, mylars, faded photocopies), or programs that need consistent QA artifacts across an entire archive.

## Frequently Asked Questions

### What is AI georeferencing?
AI georeferencing is the automatic alignment of a scanned map, aerial photograph, or legacy plan to a real-world coordinate system. Instead of a GIS technician manually identifying ground control points and matching them to a reference dataset, a machine-learning model detects features on the source image, matches them against a basemap, and computes the transformation that places every pixel at its correct geographic position.

### Do I need to pick ground control points manually with Monarcha?
No. Monarcha detects and matches ground control points automatically. Manual review and override is available for quality assurance on sensitive jobs, but the default workflow does not require any manual point picking.

### What accuracy can I expect from Monarcha?
On modern basemaps with typical aerial, topographic, or cadastral input, Monarcha achieves sub-meter accuracy. Accuracy varies with the quality of the source scan, the age of the imagery, and the density of identifiable features. Every output ships with per-point residuals so the result can be audited before acceptance.

### What coordinate reference systems does Monarcha support?
Monarcha supports more than 4,000 EPSG coordinate reference systems, including UTM, state plane, custom local mine grids, and historical national grids. The system identifies the projection automatically from marginal notations, grid ticks, and metadata when present, and falls back to user selection when the map gives no hints.

### What file formats does Monarcha accept and produce?
Input: TIFF, GeoTIFF, PDF, JPEG, and PNG. Output: GeoTIFF, Cloud Optimized GeoTIFF (COG), and other georeferenced raster formats with full CRS metadata preserved. Results drop directly into ArcGIS, QGIS, and any GIS that reads standard georeferenced rasters.

### Can Monarcha process archives at scale?
Yes. Monarcha is built for batch runs of thousands of sheets in a single job. Customers use it to process entire county archives, mineral exploration libraries, and historical aerial collections that would take a manual team months to handle.

### Is on-premises or private cloud deployment available?
Yes. Monarcha offers private cloud and on-premises deployments for government customers and any organization with data sovereignty or sensitivity requirements. The platform is SOC 2 compliant.

### How does Monarcha integrate with ArcGIS and QGIS?
Monarcha exports standard GeoTIFF and Cloud Optimized GeoTIFF files that load natively into ArcGIS, QGIS, Leapfrog, and any GIS that supports georeferenced rasters. No proprietary format, no special importer required.

## Industry Solutions

### Government
Federal, state, and local government agencies use Monarcha to digitize historical records including aerial imagery, satellite data, right-of-way (ROW) maps, parcel boundaries, deeds, and as-built documents. Applications include land management, environmental compliance, infrastructure planning, and declassified imagery processing.

### Mining & Resources
Mining and exploration companies use Monarcha to convert decades of legacy mine plans, geological maps, drill logs, and geochemistry reports into structured spatial data. This accelerates mineral exploration, mine planning, environmental remediation, and regulatory compliance.

### Land & Zoning
Data center developers, energy companies, sales tax assessors, and real estate technology companies use Monarcha to georeference and digitize zoning maps, land use plans, and cadastral records. This powers site selection, due diligence, and regulatory analysis workflows.

### Civil Engineering
Civil engineering firms use Monarcha to process subdivision plats, survey maps, infrastructure plans, and as-built drawings. This supports project planning, utility coordination, and asset management.

## Technical Information

- Deployment: Cloud-hosted SaaS, private cloud, and on-premises options
- API: RESTful API for programmatic access
- Integrations: ArcGIS, QGIS, AutoCAD, and custom workflows
- Security: SOC 2 compliance, encrypted data at rest and in transit
- Input formats: TIFF, GeoTIFF, PDF, JPEG, PNG
- Output formats: GeoTIFF, COG, Shapefile, GeoJSON, KML, GeoPackage, CSV

## Featured Blog Posts (Georeferencing)

- [Heterogeneous vs Homogeneous Georeferencing](https://monarcha.ai/blog/heterogeneous-vs-homogeneous-georeferencing): The modality gap and why traditional georeferencing tools fail on hand-drawn plats, mylars, and faded zoning maps.
- [QGIS Georeferencer vs Automated AI Georeferencing](https://monarcha.ai/blog/qgis-georeferencer-vs-automated-ai): When the free QGIS Georeferencer plugin is the right call and when automated AI georeferencing earns its cost.
- [How to Georeference a Scanned Map: A 2026 Guide](https://monarcha.ai/blog/how-to-georeference-a-scanned-map): Step-by-step tutorial covering UTM, state plane, and custom grids.
- [Introducing the Survey Map Georeferencer](https://monarcha.ai/blog/survey-map-georeferencer-launch): AI georeferencing engine trained on 100,000 annotated map pairs for plats, surveys, and field sketches.
- [Soviet Military Maps of China, Georeferenced and Open](https://monarcha.ai/blog/soviet-military-maps-china): Cold War-era Soviet military topographic maps covering China.
- [What is Georeferencing?](https://monarcha.ai/blog/what-is-georeferencing): Background primer on georeferencing and how AI changes the workflow.

## Contact

- Website: https://monarcha.ai
- Email: founders@monarcha.ai
- Request a demo: https://cal.com/james-monarcha/monarcha-enterprise