Mineral-scale prediction

Rigorous scale prediction for your brines — in your AI workflow, inside your network

Crysalt predicts which minerals scale, and how much, as a brine goes from reservoir to surface — barite, calcite, sulfides and more. Run it yourself, so the AI assistant you already use can call it on demand, and your production data never leaves your environment.

Pitzer-grade · validated vs PHREEQC · works with your AI tools via MCP · runs in your network

In your AI assistant

“Will this produced water scale at 60 °C and 50 bar?”

Crysalt answers locally — your data never leaves your network
Scale prediction 60 °C · 50 bar converged
2 minerals supersaturated — scaling predicted
BaSO4420 mg/L
SR 252 · SI 2.40
CaCO3200 mg/L
SR 96 · SI 1.98
Full speciation · activity coefficients · Ksp · ionic strength
10Oilfield scale minerals
21Aqueous species tracked
200+bar — T,P-corrected Ksp
Pitzer · Harvie–Møller–Weare Validated vs PHREEQC Local MCP server + REST API

Why it costs to guess

Mineral scale plugs tubing, chokes, and topside equipment — driving deferred production, intervention costs, and integrity risk. Predicting where and how much scale forms as a brine depressurises and cools is the difference between managing it and reacting to it — and with Crysalt you can do it without that data ever leaving your hands.

Runs where your data lives

Your brine analyses are sensitive — they shouldn’t have to leave for someone else’s cloud just to get a number back. Crysalt runs as a server inside your own environment: your laptop, a workstation, or your company’s network. The AI assistant you already work with can call it on demand — run a prediction, cite the numbers, act on them — and nothing you enter goes anywhere else. It connects through MCP (Model Context Protocol), the open standard your AI tools already speak. Set it up from the docs in minutes, or hand them to your IT team.

Download the server
Runs onYour machine or network
Connects viaMCP · REST API
Your dataNever leaves
SetupFrom the docs

Built for production decisions

Rigorous

Validated Pitzer ion-interaction geochemistry for concentrated oilfield brines — not black-box correlations.

The scales that matter

Barite, calcite, gypsum, anhydrite, celestite, siderite, halite, and the sulfides — FeS, ZnS, PbS.

Reservoir to surface

Temperature, pressure, pH, CO₂/H₂S gas partitioning, and water mixing — a single point or a full P,T profile.

Transparent

Surfaces speciation, activity coefficients, convergence, and the limits — so you, and your AI, can trust the number.

The science, in the open

Crysalt computes aqueous equilibrium and mineral precipitation with a Newton-Raphson solver over mass-balance equations, using the Pitzer activity model for accurate saturation ratios in high-salinity brines, with pressure-corrected solubility products. The methodology, the parameter provenance, and the scope are published in full — including what the engine does not yet model.

Read the methodology
Activity modelPitzer (Harvie-Møller-Weare)
SolverNewton-Raphson equilibrium
SolubilityT,P-corrected Ksp
Gas phaseCO₂ / H₂S / CH₄ partitioning
InterfaceLocal MCP server + REST API

Predict scale, privately, in your AI workflow.

Try it in your browser — then run it in your own environment.

Contact

About this project

Crysalt is an independent research project, built and maintained by one person — a petroleum engineer whose doctorate was in exactly this: the thermodynamic modelling of scale precipitation. It is free to download and use internally under its licence. No forms, no tracking, no mailing list.

  • Questions about the science — methodology, parameter provenance, or where the model stops.
  • Running it yourself — the docs cover setup; ask if something is unclear.
  • Rights — commercial, redistribution, hosting and OEM rights are reserved. Enquiries welcome.

Use of this site and the Playground is subject to the Terms of Use.

An unhandled error has occurred. Reload 🗙