Vexcel Holding Corporation announced on July 1, 2026, in Centennial, Colorado, that it has developed a Model Context Protocol (MCP) server, bringing its industry-leading high-resolution aerial imagery and geospatial intelligence directly to AI assistants like Microsoft Copilot and ChatGPT. Licensed customers can now query and retrieve Vexcel’s massive library of orthoimagery, 3D models, and geospatial data through natural language, all without leaving their AI environment. The move marks the first time a major geospatial data provider has bridged its proprietary datasets with the rapidly expanding MCP ecosystem, effectively turning conversational AI into a powerful remote-sensing workstation.

The Vexcel Advantage: Why Its Imagery Matters

Vexcel is not a household name, but in the world of photogrammetry and remote sensing, it is a titan. The company operates one of the largest aerial survey programs on the planet, collecting over six million square kilometers of imagery annually across the United States, Europe, and Australia. Its UltraCam and Condor sensors capture orthophotos with a ground sample distance as fine as 5 cm per pixel—enough to see individual manhole covers and lane markings. This precision has long made Vexcel the go-to source for insurance carriers assessing roof damage, city planners mapping urban heat islands, and construction firms monitoring site progress. Until now, accessing that imagery meant logging into a dedicated Vexcel web viewer, using desktop GIS software, or wrangling APIs—none of which fit well into the conversational workflows emerging around large language models.

What Is the Model Context Protocol?

The Model Context Protocol, originally open-sourced by Anthropic in 2024, defines a standardized way for AI models to interact with external tools and data sources. Think of it as a USB-C for AI: a universal connector that lets an LLM securely plug into a remote server, ask for information, and receive structured responses. MCP servers can expose anything from a company’s internal knowledge base to a live weather feed. For Vexcel, the protocol means its geospatial APIs can be wrapped in an MCP server that handles authentication, rate limiting, and query translation, so that a prompt like “Show me the most recent aerial of this property” in Copilot or ChatGPT becomes a fetch request to Vexcel’s image tile service.

Crucially, MCP is both model- and platform-agnostic. Any MCP-compatible host—whether Microsoft’s Copilot Studio, OpenAI’s ChatGPT plug-in framework, or a custom-built assistant—can connect to Vexcel’s server once a customer provides an API key or OAuth token. Vexcel’s implementation follows the protocol’s latest 2025 specification, adding geospatial-specific capabilities such as bounding-box queries, coordinate reference system negotiation, and time-range filters. The server returns a combination of image URLs, metadata JSON, and vector outlines, all of which the host AI can interpret, display, or chain into further analysis.

How It Works Inside Copilot and ChatGPT

The user experience is designed to be seamless. A licensed Vexcel subscriber installs the Vexcel MCP plugin from the Copilot or ChatGPT marketplace, enters their credentials, and immediately gains access to their organization’s imagery layers. In Microsoft 365 Copilot, for example, a field adjuster might type: “Give me the post-storm imagery for the five houses on Oak Street from the July 3rd flight,” and Copilot will fetch the corresponding orthomosaic tiles, overlay property boundaries from the user’s own CRM, and present a before-and-after comparison in a sidebar. In ChatGPT, a researcher could ask, “What was the average greenness index of this farm field in June 2025 and 2026?” and get not only the numbers but a side-by-side map with NDVI values derived from Vexcel’s multispectral layers.

Behind the scenes, the AI assistant breaks down the natural-language request into a series of MCP tool calls. For a bounding-box query, it might first invoke a geocoding tool to convert an address into coordinates, then call Vexcel’s search_imagery function with those coordinates, a desired date range, and a resolution preference. The MCP server validates the license, checks availability, and returns a set of georeferenced image URLs along with metadata like capture date, sun angle, and cloud cover percentage. The assistant can then use its native rendering capabilities or a lightweight mapping component to display the results. Because all processing stays within the MCP session, the user never leaves the conversation.

Who Gets Access and What It Costs

Vexcel’s MCP server is not a free-for-all. Access is tied to an active Vexcel license, which typically includes imagery subscriptions, data storage, and usage limits. Organizations already paying for Vexcel’s online mapping or API services will be able to enable the MCP endpoint with minimal additional configuration. New prospects can request trial access through a program called “Vexcel Spatial AI Lab,” which offers a limited imagery footprint and capped monthly queries for proof-of-concept testing. Vexcel has not disclosed specific pricing tiers for the MCP add-on, but industry analysts expect it to follow the company’s existing volume-based model, which starts at a few hundred dollars per month for small footprints and scales into five-figure annual contracts for nationwide coverage.

Real-World Applications: From Insurance to Urban Planning

The union of high-resolution aerial imagery and conversational AI opens doors across dozens of industries. Here are a few concrete scenarios that become possible—or dramatically simpler—with the Vexcel MCP:

  • Property insurance underwriting and claims: Adjusters can query “Show me the property at 123 Main St. as it looked in March 2024, October 2025, and last week,” getting a visual timeline of changes. AI models can then flag new structures, missing sections of roofing, or vegetation encroachment, all through dialog.
  • Real estate portfolio management: Commercial real estate managers can ask, “Which of my properties in Phoenix have any visible roof damage or ponding water?” and receive a ranked list with annotated images, pulling from Vexcel’s archive that spans multiple years.
  • Agriculture and forestry: Agronomists can request “Create a vigor map for Section 14 this season and compare it to last year,” leveraging Vexcel’s 4-band imagery (RGB + near-infrared) to compute vegetation indices on the fly.
  • Disaster response and resilience: Emergency managers can connect Vexcel’s MCP to a chat interface and ask, “What did the flooded area look like yesterday and before the levee broke?” The AI fetches the most current post-disaster imagery captured by Vexcel aircraft often within 24 hours of an event.
  • Municipal planning and code enforcement: City inspectors can chat, “Show me all backyard pools and unpermitted structures in this neighborhood,” and have the AI overlay parcel boundaries and highlight discrepancies against building permits.

Technical Edge: Speed, Resolution, and Freshness

What makes Vexcel’s deployment particularly compelling is the underlying data quality. Vexcel refreshes its urban and suburban imagery multiple times per year, ensuring that what users see is rarely more than a few months old. The server can stream image tiles at full resolution, allowing AI assistants to zoom into detail as if using a desktop GIS. Moreover, because the MCP server sits on Vexcel’s own infrastructure—not a third-party aggregator—latency is low and data sovereignty is maintained. Large enterprise customers who already trusted Vexcel with petabytes of imagery can simply expose a subset to their AI tools without risky data migration.

The Competitive Landscape

Vexcel’s move comes at a time when geospatial data is becoming table stakes for AI assistants. Microsoft has its own Azure Maps and Bing Maps imagery, while Google offers Earth Engine and Maps APIs. However, the resolution of commercial-grade mapping APIs typically tops out at around 15 cm per pixel, and the imagery refresh cycles can be seasonal at best. Vexcel’s 5 cm imagery and frequent urban recaptures give it a noticeable advantage in professional use cases where precision matters. Other aerial survey companies like Nearmap and EagleView offer high-res imagery, but neither has announced MCP support as of this writing. Vexcel’s first-mover position could entrench it as the default geospatial backend for enterprise AI workflows, much as Snowflake became the data warehouse for business intelligence.

Privacy and Ethical Considerations

High-resolution imagery that can be summoned with a simple chat message raises legitimate privacy concerns. Vexcel has addressed this by enforcing its existing strict usage controls through the MCP gateway. Only authorized users with a contractual license can access the data, and all queries are logged for audit trails. The system uses the same role-based access controls that govern its direct API—so a municipal user might see only public infrastructure layers, while an insurer gets property-level detail but only for addresses directly tied to an active policy. Additionally, Vexcel’s imagery is captured from aircraft at altitudes above a threshold that, in most jurisdictions, complies with privacy regulations, and faces are not identifiable at the resolutions provided to bulk customers. Nonetheless, Vexcel has committed to a review board that will evaluate new use cases as they emerge, particularly those involving AI-driven surveillance or profiling.

Developer Ecosystem and Customization

Beyond the out-of-the-box plug-ins for Copilot and ChatGPT, Vexcel’s MCP server is open for any developer to integrate. Because MCP is an open protocol, teams can build custom AI agents that combine Vexcel’s imagery with other enterprise data sources—say, a construction progress bot that fuses Vexcel orthophotos with drone captures from a DJI dock, or a environmental compliance assistant that cross-references aerial imagery with regulatory maps. Vexcel has published a Python client library and a set of sample prompts on its developer portal to accelerate adoption. The company also hosted a pre-launch hackathon in partnership with Microsoft, which produced a prototype for a farm-equity valuation tool that reduced manual appraisal time by 60%.

What’s Next: 3D, Oblique, and AI-Powered Analysis

Steven MacLean, Vexcel’s CTO, hinted that the MCP server will soon support not only orthophotos but also oblique views and 3D point clouds derived from the company’s recent push into full-city lidar and mesh models. A future update will enable queries like, “Show me the southeast-facing facades of all buildings on this block,” pulling from Vexcel’s Oblique imagery. Even more ambitious is the integration of Vexcel’s own AI analytics directly into the MCP workflow. The server could expose functions such as detect_changes, classify_land_use, or measure_roof_area, allowing the host assistant to trigger these compute-heavy jobs without the user writing a single API call. That vision transforms Copilot from a simple image retriever into a full geospatial analyst that understands context and follows complex, multi-step instructions.

The Bottom Line

Vexcel’s MCP server announcement is more than a new product feature; it is a glimpse into a future where high-fidelity, real-world data becomes as accessible through AI chat as any database table or document. By choosing the open Model Context Protocol, Vexcel ensures its imagery can flow into the tools users already use every day, rather than requiring them to learn yet another specialized platform. For Windows users embedded in the Microsoft ecosystem, the immediate availability of Vexcel’s imagery inside Copilot could supercharge everyday tasks in Excel, Teams, and Power BI, turning a simple spreadsheet of property addresses into a rich, visual dashboard. The geospatial AI race is officially on, and Vexcel just planted a very high-resolution flag.