AI and integrations
AI-native, from the data model up.
Agents query and edit your data through MCP, and every change lands as a draft for an editor to review. Integrations and webhooks put the result where your apps need it.
For AI agents
An AI teammate for your map data.
Categorize, geocode, draft descriptions, flag duplicates and translate. Nothing is published until an editor approves it.
Spatibase AIWORKING
✦CategorizedTrails, campsites, viewpoints1,204
⌖GeocodedFrom street addresses96
✎Descriptions draftedIn your editorial tone214
⚑Duplicates flaggedWithin 25 m, similar names12
⇄TranslatedEN → DE, FR, ES642
✓Ready for reviewNothing is published automatically38
Ask your datavia MCP
Which campsites near water have no description?
Found 34 campsites within 500 m of water with an empty description. I drafted all of them from their location, type and nearby features.
34 drafts0 publishedReview changes →
Agents
Connect your own agents
An MCP server exposes your project's schema and content to agents, scoped by what you allow. Agents query and edit; they cannot publish on their own.
- Scoped, token-based access
- Every edit lands as a draft
- Every change is in the audit log
{
"mcpServers": {
"spatibase": {
"url": "https://mcp.spatibase.com/projects/{id}",
"headers": { "Authorization": "Bearer <token>" }
}
}
}Fits the stack you have.
For developers
MapLibre / Mapbox
Vector tiles and GeoJSON, ready to use.
For developers
QGIS
Use published GeoJSON and vector tiles in desktop GIS.
For developers
Webhooks
Trigger workflows when content is published.
Build your spatial backend
Start free, or talk to us about your organization.