The Petropt Platform

One platform. Three ways to work.

Tools, Agents and Enterprise — built on the same petroleum-engineering primitives. Use what you need today, scale into the rest.

Try the Tools → Request a Briefing

How the pieces fit.

Three product surfaces, one shared core. The open-source layer is auditable; the layers above use it.

PETROPT TOOLS Public, self-serve tools.petropt.com 169 free + 141 Pro tools PETROPT AGENTS Scheduled automation DCA · production close Anomaly watch · reserves QC PETROPT ENTERPRISE Private deployment Shared workspace · SSO On-prem · private VPC SHARED CORE petropt petro-mcp open-source · MIT

Surface 1

Petropt Tools

tools.petropt.com

169 free tools + 141 Pro tools across 10 engineering suites — reservoir, production operations, drilling, completions, petrophysics, well testing, facilities, flow assurance, artificial lift, economics.

Try without a login. Save runs with one. Every output shows its method, its assumptions, and its citation: Arps (1945) for exponential and hyperbolic decline, Fetkovich (1980) for boundary-dominated flow, Duong (2011) and SEPD for unconventional, Glaso (1980) for PVT, Hagedorn-Brown for multiphase lift. The math is on the page.

Open the tools →

Surface 2

Petropt Agents

petropt.com/agents

Pre-built agents — DCA agent, daily production reporter, anomaly watcher — plus a registry to compose your own.

Agents are scheduled or triggered. They call the same MCP tools we publish open-source, on your data, and write the result into your engineering systems. Not a black box: every agent run shows its tool calls and intermediate state.

See the agents →

Surface 3

Petropt Enterprise

Founder-led engagements · by demo only

Private deployment with the shared engineering workspace, plus on-prem option, SSO and custom workflows — for operators with data-residency requirements and engineering teams that need the workflow in their own VPC.

Inside the private deployment, the shared workspace persists every run with its inputs, equations and outputs — re-open last quarter's DCA in one click, share results URLs across the team, and reproduce any forecast a year later with the same parameter choices it was built from. Multi-well, multi-pipeline, multi-artifact.

SAML / OIDC. Full audit log on every workspace action. Custom workflow development. Founder-led engagements, with a hand-picked network of senior engineers and specialists brought in per scope. Access is by formal demo request.

See the enterprise model →

Field Workflows

What this looks like inside a real engineering team.

Three workflows we build for operators, each tying Tools, Agents and the Enterprise workspace into one loop.

AI-Assisted Decline-Curve Analysis

A reservoir team running 800 wells across two basins refreshes decline forecasts monthly. The DCA agent reads production from the workspace, fits Arps exponential, hyperbolic, harmonic, and Duong models with constrained b-factors, flags fits with R² below threshold or rate jumps that indicate workover events, and routes those exceptions to the on-call reservoir engineer. The engineer reviews the queue, accepts or refits, and the workspace persists the final P10/P50/P90 with every assumption attached.

9 days → 1.5 days monthly DCA cycle.

Representative outcomes; engagements under NDA.

Daily Production Report Agent

A production engineer used to spend Monday mornings rebuilding the daily report from SCADA exports and allocation spreadsheets. The agent runs at 5am: pulls allocated rates, downtime codes, paraffin and water-cut flags from the workspace, generates the daily exception report, and posts it to Slack and to the operations folder before the morning meeting. Exceptions route by area; routine wells stay routine.

~12 hours / engineer / week recovered.

Representative outcomes; engagements under NDA.

Parameterized Well Economics for Proposals

An A&D team running diligence on 1,200 wells across three packages used to rebuild NPV, payout, and breakeven in Excel for each scenario. The workspace template ingests the production forecast, applies a versioned price deck, severance tax, and working-interest assumptions, and outputs NPV/IRR/payout in batch — with one-click diff against the previous run when the partner pushes back on a number.

3× proposal throughput per senior engineer.

Representative outcomes; engagements under NDA.

Open Core

Inspect the math. Run it locally. No vendor lock-in.

The reservoir, production, and economics methods that drive everything else are open-source Python. Read the implementation. Pin a version. Use it without us.

petropt

The petroleum engineering Python library. 50+ validated functions across PVT, decline, IPR/TPR, petrophysics, drilling, production engineering. MIT-licensed.

PyPI version PyPI downloads MIT License
pip install petropt View on PyPI →

petro-mcp

The MCP server that exposes petropt to AI assistants — 70 petroleum engineering tools callable from Claude, ChatGPT, or any MCP-compatible client.

PyPI version PyPI downloads MIT License
pip install petro-mcp Product page →

One contract. One data plane.

Your data lives in one place. Surface changes — a new tool, a new agent, an enterprise add-on — don't break your audit trail. Every result on every surface ties back to the same versioned inputs, the same methods, and the same workspace history.

That's the difference between a platform and a folder of point tools.

Use what fits today.

Free tools to try the math. Agents for the work you don't want to do every Monday. Enterprise when you need it private — with the shared workspace, SSO and custom workflows built in.

Open the Tools → Request a Briefing