Skip to content
Tresca

Leapfrog AI integrations

No MCP server or agent skill for Leapfrog is listed yet (checked 4 Nov 2026). No MCP server or agent skill targets Leapfrog Geo itself yet. The practical route for an agent is Seequent Evo: Leapfrog publishes selected objects and block models to Evo, and Seequent's own Evo MCP server and Python SDK work on that cloud data, not on the open desktop project. Directory entries called "Leapfrog MCP" belong to an unrelated browser-automation product, so check the author before you install one.

Bentley (Seequent)None listed yetChecked 4 Nov 2026

How AI connects to Leapfrog

An MCP server or skill can only do what Leapfrog lets programs do. Leapfrog exposes 2 automation interfaces:

  • Seequent Evo APIs and evo-python-sdk (Python): Leapfrog Geo has no documented desktop scripting interface; automation goes through Evo instead. The evo-python-sdk on GitHub bundles clients for the Evo File, Geoscience Object, Block Model, Colormap and Compute Tasks APIs, installed together as the evo-sdk package. Every call needs an app registered in the Seequent Developer Portal, which in turn needs a Bentley developer account, and an Evo access token tied to your organisation, hub and workspace. Documentation
  • Publishing from Leapfrog Geo to Evo (Other): The bridge between the desktop modeller and anything an agent can reach is the Evo menu in Leapfrog Geo 2026.1. Only some project objects can be published, an object published to one workspace can't be sent to another, and folder names are fixed after publishing. Regular, octree and fully sub-blocked models publish; variable Z sub-blocked models do not. Updated Evo objects are reloaded into the project only when you choose. Documentation

What Bentley (Seequent) offers

Seequent publishes an Evo MCP server (SeequentEvo/evo-mcp, labelled early development) that exposes workspaces, geoscience objects and block models to MCP clients, but it reads Evo cloud data rather than a running Leapfrog session. Bentley's MCP page lists STAAD, MicroStation and other infrastructure products and no Seequent product. Seequent's PDAC 2026 programme pitched machine learning for structural interpretation and Evo as the base for AI workflows.

Checked 4 Nov 2026. Sources: SeequentEvo/evo-mcp README, Bentley MCP servers, Seequent brings AI and cloud-enabled geoscience to PDAC 2026 (Bentley news, 11 February 2026)

What an integration for Leapfrog could do

  • Pull a block model that a geologist published from Leapfrog Geo into Evo and report grade and tonnage by domain without opening the desktop project.
  • Load collar, survey and interval CSVs as Evo downhole collections so they can be brought back into Leapfrog Geo, flagging schema problems before import.
  • Compare two published versions of a geological model object in an Evo workspace and summarise what changed between revisions.
  • Answer questions about Evo publishing rules, such as which sub-block model types can be sent from Leapfrog and why a published folder can't be renamed.

Know of one, or building one?

If an MCP server or agent skill for Leapfrog exists, tell us and we'll check it and list it. You can also ask for one, which helps builders see demand.

Submit a listing Request an integration

Tools with AI integrations for similar work

MicromineMicromine (Weir)None listed yetVulcanMaptek1 serverDatamineDatamine1 serverPLAXIS 2D / 3DBentley (Seequent)2 servers · 1 skillGeoStudioBentley (Seequent)1 serverMicroStationBentley2 servers · 1 skillQGISOpen source2 servers · 3 skillsArcGISEsri4 servers

Questions

Is the "Leapfrog MCP" server in MCP directories for Seequent Leapfrog?

No. The Leapfrog entry listed on MCP directory sites such as mcpservers.org is a browser-automation tool that shares the name. It has nothing to do with Seequent's geological modelling software and can't open a Leapfrog Geo project.

Can Claude build or edit a Leapfrog Geo model?

Not directly, as of 4 November 2026. Leapfrog Geo exposes no scripting interface for the desktop project, so an agent can only work with objects you publish to Seequent Evo, through the Evo APIs or Seequent's Evo MCP server.

Does Seequent have an official MCP server?

Seequent publishes an Evo MCP server on its SeequentEvo GitHub organisation, described as early development. It manages Evo workspaces, objects and block models; it does not control Leapfrog Geo, and Bentley's own MCP page doesn't list it.

Do I need a licence to script against Evo?

You need access to Seequent Evo and an Evo app registered in the Seequent Developer Portal, which requires a Bentley developer account. Without Evo, there is no supported programmatic way into Leapfrog data beyond file export.

AI-generated models, drawings and results must be checked by a qualified engineer before use.

Sources (checked 4 Nov 2026): Seequent Developer Portal · SeequentEvo/evo-python-sdk README · SeequentEvo/evo-mcp README · Publishing data to Seequent Evo (Leapfrog Geo 2026.1 help) · Bentley MCP servers · Seequent brings AI and cloud-enabled geoscience to PDAC 2026 · Leapfrog product page

Leapfrog is a trademark of Bentley (Seequent). Tresca is independent and not affiliated with or endorsed by Bentley (Seequent).