Q4 2025
New features and improvements introduced from October to December 2025
Unified access to Data Observatory subscriptions and improved subscription management
November 18th, 2025
Improvement Workspace
With this release, we’re making it simpler and more consistent for users to access and work with data from their Data Observatory subscriptions. Access to data has now been fully unified to always be via your own data warehouse connections. Additionally we've also improved the way Admin users can manage the organization's Data Observatory subscriptions from the Settings section in the CARTO Workspace.
We’ve unified access to the data from Data Observatory subscriptions to always be rom the end-user data warehouse connections. As announced earlier this year, we have deprecated the Data Observatory tab in Data Explorer, Builder, and Workflows. This tab previously exposed subscriptions only through a small set of connections (i.e. CARTO Data Warehouse and BigQuery US multi-region). Since all subscriptions are now available directly via data warehouse connections, the tab has been removed to avoid confusion.
The Data Observatory section in Settings has been significantly improved. It now serves as the central place to manage your organization’s subscriptions, showing to which data warehouse each subscription has been transferred, and allowing users to request new transfers so the data is available directly in their data warehouses.

Starring data assets for quicker access
November 18th, 2025
Improvement Workspace
Users are now able to star items at any level in the Data Explorer, including connections, projects/databases/schemas, and the data tables themselves. Simply click on the star icon next to any item in the Data Explorer and then use the Starred only filter to show just your starred items.
This is especially helpful to users that have connections or data assets that are recurrently used in their maps and workflows. No more browsing the data tree until you find what you need!
Your starred items are now also easily accessible from the "Add data source" flow in CARTO Builder and from the data sources panel in CARTO Workflows.
To learn more about starring items and the Data Explorer in general, check out our documentation.

Integrate CARTO maps anywhere with the new authenticated embeds and reactive map events
November 7th, 2025
Improvement Builder
Embedding maps from CARTO in other webpages and applications just became exponentially easier and more powerful thanks to two additions to our platform:
New methods for seamless and secure private embedding: We added two new strategies to embed private maps securely, without having to publish the map or forcing the users to login in a different tab or browser. Developers can also re-use existing authorization in their applications. Learn more about private embedding strategies.
Build bi-directional interactive experiences with our embedded events: Embedded maps from CARTO now send
postMessageevents every time something changes in the map. This allows the parent application to react, creating bi-directional interactive experiences when combined with our embed URL parameters. Learn more about embedding events.
We're excited to see where you will embed your next CARTO map!

New map interaction tools for AI Agents
November 6th, 2025
New Builder
AI Agents can now interact directly with your maps through two new tools:
Dynamic marker placement: Ask the AI Agent to mark specific locations, and it will instantly place markers on your map. Simply provide an address, place name, or coordinates—the agent handles geocoding and placement automatically.
Spatial filtering by area: The AI Agent can define custom areas of interest to filter your data dynamically. When an area is set, all map widgets and layers update automatically to show only data within that region.
These tools enable your AI Agent to provide immediate visual context and perform focused analysis on specific geographic areas without manual configuration.
Learn more in our documentation.
Introducing Agentic GIS in CARTO: making spatial insights available to everyone
October 8th, 2025
New Builder
We are incredibly excited to announce new features that bring enterprise-grade geospatial agentic experiences to CARTO.
Introducing AI Agents in Builder: CARTO AI Agents (now in General Availability) provide a conversational interface in your maps where your end users can get instant and actionable geospatial insights through natural language.
AI Agents can now query sources, generate layers and more: We've added a ton of exciting capabilities that allow agents to reason and perform geospatial analysis autonomously.
Integrate Workflows as tools for your AI Agents: From building operational dashboards to running complex analyses, your AI Agent can be supercharged with your own custom workflows enabled as MCP tools.
Evolved experience to tailor your Agent: You can now reference tools, sources, and other context available in the map when customizing your agent.
Use your own AI models: Configure your own AI models and maintain total control over the AI technology used. Supported providers include Google Gemini and Open AI, with others coming soon.
With CARTO you can now create and share access to powerful geospatial AI Agents tailored to your specific needs. Combine your custom prompt instructions with CARTO's built-in geospatial intelligence and your own workflows, and build trustworthy AI solutions that make complex geospatial analysis accessible to any user within your organization.
Get started today by enabling CARTO AI in your organization.
And learn more about Agentic GIS in our announcement blog post!

Turn your AI Agents into geospatial experts with CARTO MCP Server
October 8th, 2025
New Workflows
CARTO now supports the Model Context Protocol (MCP), a standard that enables AI Agents to interact with external tools and data sources. With the new CARTO MCP Server, organizations can now expose their own geospatial Workflows as MCP Tools that any MCP-compliant agent can use.
This release allows GIS teams to design custom workflows in CARTO—defining inputs, outputs, and logic specific to their spatial problems—and make them available to AI Agents through the MCP Server. Each tool includes detailed metadata following the MCP specification, ensuring interoperability across agentic AI environments.
By combining Workflows and the MCP Server, organizations can empower AI Agents to perform advanced spatial analysis, automate geospatial decision-making, and connect AI-driven applications to their cloud data infrastructure.
Last updated
Was this helpful?
