Petrel MCP servers and AI agent skills

How AI connects to Petrel

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

• Ocean software development framework (.NET): Ocean is SLB's framework for building plug-ins that run inside Petrel, Techlog and Studio. It is built on Microsoft technology, and the Ocean for Petrel developer guide on SLB Developer walks through creating a plug-in project, registering commands and accessing domain objects such as seismic, surfaces and grids. Plug-ins are sold through the Ocean Store, which lists more than 120 for Petrel. An MCP server for Petrel would most likely be an Ocean plug-in. Documentation
• Python Tool Pro for Petrel (Python): Python Tool Pro, built with SLB's partner Cegal, is a module for Petrel 2024.x that runs Python scripts directly on project data: wells, logs, seismic cubes, surfaces, grids and interpretations. Results are written back to the project in one call, and SLB says it can scale from desktop to HPC. It is the closest thing to a scripting API an external agent could call, but it needs the module licence. Documentation

What SLB offers

SLB launched Tela, an agentic AI assistant built on its Lumi data and AI platform, on 3 November 2025. Tela in Petrel (Subsurface Insights and Reservoir Insights, catalogue pages dated 29 June 2026) answers natural-language questions about project data, retrieves Petrel and simulator documentation, and interrogates simulation ensembles. It runs inside SLB's software, not as an MCP server for outside clients. SLB's separate Developer MCP server searches developer docs only.

Checked 4 Nov 2026. Sources: SLB unveils Tela agentic AI (press release, 3 November 2025), Tela in Petrel, Subsurface Insights (SLB Tela catalogue), Tela in Petrel, Reservoir Insights (SLB Tela catalogue), Petrel AI assistant (SLB video page), AI Assistant and MCP integration (SLB Developer)

What you can do with Petrel and AI

• List the wells, logs, surfaces and grids in an open Petrel project and summarise what each interpretation folder contains before a review.
• Run a log-QC or facies-prediction script through Python Tool Pro and write the result back as a new log, then report what changed.
• Pull simulation results attached to a Petrel case into a comparison table across an ensemble of realisations.
• Help a developer scaffold an Ocean plug-in by looking up the right domain objects and command registration in the Ocean for Petrel guide through the SLB Developer MCP server.
• Export horizons or well tops in a format another tool can read, such as an ECLIPSE grid include or a GIS layer.

Taken from the listings' own documentation. Check each listing for exactly what it supports.

Installing

The listings here document setup for ChatGPT, Claude, Codex, Cursor and VS Code. Install steps, copied from each project's own documentation, are on each listing's page. Most Petrel integrations run on your own computer next to Petrel, so you need Petrel installed and licensed on the same machine unless the listing says otherwise.