Welcome to Ophiolite

Ophiolite builds connected scientific workflows for subsurface data: application connections, understandable revisions, bounded calculations and independent reading.

This documentation covers concepts and the current pilot scope. An operator installs Ophiolite on one Linux host from the server image (Install); Python applications connect with the ophiolite package (Ophiolite CLI).

• Check current availability.
• Understand follow and pin.
• Read a small dataset independently.
• Explore connector capabilities.

Return to the product website.

Own the data and the work behind it

• How the architecture fits covers the seven design diagrams.
• Scientific lineage and AI datasets explains how exact inputs, operations and labels support reproducible training.
• Lineage, audit, telemetry and analytics explains customer visibility and separate controls.
• Publication and recovery explains the reliability contract.

For end users

Open a local LAS, open a shared asset and publish a selected curve. These are separate development-workbench actions; local opening never uploads data.

Understand coordinates and CRS conversion, use and troubleshoot QGIS live maps, or open the public USGS fixture without an Ophiolite account.

For developers

Component boundaries and the standalone numerical library explain what belongs in the client, connectors, platform and execution runtime.

Read the scalar-map data contract, publication and notification APIs, and Python tools / SDK status. These references describe the qualified pilot and repository interfaces; a packaged public SDK remains planned.

Manage effective project access and inspect exact scientific assets.

QGIS account-based browsing

The private QGIS pilot now supports independent browser-approved sign-in, authorized project/asset browsing, follow/pin/local-copy selection and encrypted credential storage. See QGIS Workspace guide and connector identity API.

Inspect operation history and application observations or use the activity/export API.

• Understand source qualification and try the synthetic independent-reader demo.

• Connect your own file, SQL or OSDU source.

Edit SQL wells and subscribe to Workspace updates.

Build a curve application

Configure a LAS input in Workspace and run a standalone Python correction. The bounded pilot publishes a separate LAS result; it does not host your code.

Administer an approved PostgreSQL connection and share retained Python results.

Native OpendTect pilot

Import a GR log, edit it in OpendTect and return a separate LAS. Qualified for Linux 7.0.10 with explicit precision and missing-edge limits.

Read exact scientific assets and reuse a permitted Python result.

For a local LAS, start with self-service upload and the downloadable Python kit. This route keeps original bytes and removes server-file configuration from the supported handoff.

Read scientific data locally

Use Python to discover and download an exact LAS or curve. A permitted read needs no calculation configuration or output definition. Plot locally, or add a supported output file through the existing Connect upload flow when needed.

Use the Python and TypeScript SDK for supported 0.x clients, local workflow templates and recovery guidance over preview contracts.