Deterministic, traceable should-cost estimates for industrial projects — CAPEX, OPEX and TCO with the AACE accuracy range built in. Three levels of simulation: open references, Costonomics-calibrated cost books, your own data. No black box, no guessing.
The same scope always yields the same estimate. Every line cites its source, every assumption is numbered, every factor applied — FX, escalation, location, maturity, learning curve, material — is shown.
A size outside a correlation's validity range, a missing exchange rate, an unknown material: the engine raises an explicit error instead of inventing a number. An error is information.
Results always carry the AACE class accuracy range and contingency — and a completeness guard makes sure contingency never hides missing scope.
Versioned cost books, parametric correlations with enforced validity ranges, and structured OPEX models — maintained yearly so your estimates never go stale.
I&C and field instrumentation, chemical process (ISBL/OSBL, Turton-based), mining — open pit, underground, ISR, beneficiation, drilling, grade control — software and execution: each domain ships as a self-contained pack, deployable alone or combined.
Structured operating costs: power, fuel, maintenance, labour, consumables, G&A — reconciled component by component.
Run-to-failure to prescriptive: maintenance scales with the sensors and software actually in scope — no savings without the enablers.
From smart sensor to digital twin — vendor, open-source and hybrid architectures, build and run costs.
Per-line currencies, FX tables, location factors, index-based escalation — and a price transposer: a 50 m³ vessel bought in France in 2020 becomes its 80 m³ equivalent in the USA in 2026, every factor traced.
Lang, Hand, capacity ratio, capital intensity, SMRP maintenance band — factored methods automatically challenge every estimate.
Variant trade-offs with break-even detection, tornado sensitivity, target-drift tracking, per-line "why this price" explain — plus a per-module DtC question bank for design reviews and the IdeaTrack register: idea lifecycle, mandatory gain rationales, savings computed from the actual BOM lines, and an interactive gain/effort matrix whose selection compiles straight into a project scenario.
House formulas calibrated on your actuals live beside the reference books — the third simulation level, and it always wins. Tier3 builds it from your ERP purchase history; ParaCost fits parametric formulas on your quote returns — review-gated, fully offline.
Tick the components to consult, generate one RFQ pack per supplier — RFQ document plus Excel quote sheet, never disclosing your should-cost — then pick which returned quotes enter the BOM and measure the gap, line by line.
Don't declare your estimate class — determine it. Score your actual deliverables (P&IDs, heat & material balance, plot plans…) against RP 18R-97 logic and get the class they support, plus the exact gaps to the next one.
A light Monte Carlo where each line's uncertainty follows its pricing basis — a firm quote is tighter than an assumption — with the implied contingency, S-curve spend phasing and the engineering FTE staffing plan.
Returned quotes compared vendor by vendor against your should-cost — price and lead-time scored, recommended mix applied in one click — while every intake feeds your private observed-price history — with the competition regime behind each price, so the sole-source premium becomes an explicit figure — and contractual index formulas revise old prices traceably.
Curated cost-driving guidance per catalogue category — sizing margins, material thresholds, standardization — shown in the line inspector at the exact moment a component is chosen. Every rule cites its source; the bank is maintained without a terminal.
An article with no catalogue home (a skid on a drawing, a local service, an unmapped take-off row) is captured, not dropped: priced only with a sourced direct price, otherwise reported and the total marked a floor — then attached, priced or capitalized into your own books.
Admin, estimator, engineering, buyer, guest — personal keys, licence-bound seats, and hard server-side rules: the engineering profile never sees a vendor-selected price.
SysOpt turns coupled design choices (pipe diameter vs pumping energy, refrigerant vs compressor size) into an exhaustive bi-level search — every point priced by this engine, the winner delivered as a regular costing project.
Technical alternatives live as patches on one reference — written by the engineering team, compiled from a DtC idea selection, or generated by the gap analysis of a revised P&ID/BOM. One synthesis prices them side by side, with a significance verdict: a delta below the changed lines' uncertainty is no basis for decision. Retain merges; every decision is journaled.
The standard project WBS reconciles the whole CAPEX to the cent — and every node shows the accuracy band of its pricing bases and the finest AACE class that data can defend. A node priced at ±35% supports Class 4, never Class 3, whatever the declared class.
Every line is priced at the deepest level available to you — and each result says which level produced it, so you always know what your number rests on.
Turton / CAPCOST correlations, Camm / USGS / O'Hara mining screening models, Lang & Hand factored methods. Fully citable, free to use — and the engine behind our free public calculators.
Included in every planOur market-calibrated prices, ratios and location factors — grouped in domain packs (I&C, mining, chemical, software, execution), versioned, sourced line by line, refreshed yearly and cross-checked against public feasibility studies (NI 43-101 / JORC / S-K 1300). Secured on our servers: the engine returns results, never the coefficients.
Professional and aboveHouse formulas fitted on your actuals and your vendor quotes. Purely yours: never shared, never used to train anything, on-premise if you want it — and when present, they take precedence over every other level. With the local-LLM take-off option, even your P&IDs and EPC bills of materials never leave your network — and the Tier3 / ParaCost importers build this level from your ERP history and quote returns, review-gated and fully offline.
Platform licence & Calibration subscription
80 %
of a project's cost is committed by design decisions — taken weeks before the estimate lands on someone else's desk. Close that loop: cost feedback at the moment of choice, per discipline.
Per-discipline cost gradients ("what your choices cost" — a DN up, an alloy swap, a redundancy), standard cost curves per technology, equipment datasheets — and a Design-to-Cost question bank per catalogue module, each question pointing at the pricing mechanism that quantifies its lever. Regenerated live from the current cost books, so they never go stale.
Variant trade-off studies with TCO break-even, tornado sensitivity to know which parameters to firm up, target-drift tracking against the allocated envelope at every review — and BOM take-off: your equipment lists become a should-cost in minutes.
Close-out actuals, ERP purchase history and quote returns become your calibrated cost books and house formulas — the Tier3 importer and the ParaCost formula builder industrialize it — a level-3 cost base owned by your office, on your infrastructure, continuously measured against the market. That is what "the keys to the mechanics" means: you build yours.
Search the catalogue in English, French, German or Spanish, or start from an AACE class template — in the four-language web workshop or a plain JSON file. Set quantities (expressions allowed), sizes, pressures, materials and your industry service class; one click checks the whole project and reports every issue at once. Vendor quotes drop in per line and always win.
Adjustment chain base price × FX × escalation × location × maturity × learning × material, erection and indirects, engineering, margin — then AACE contingency and the class accuracy range. Guards flag double counts, missing scope and incoherent strategies.
One command builds the dated review dossier: full traceability report, live Excel model, Basis-of-Estimate document and tornado sensitivity — plus trade-off studies and estimate-to-estimate diffs when the design moves. At the gate, the maturity questionnaire tells you which AACE class your deliverables actually support, and the detailed analysis adds P10/P50/P90 and the S-curve cash flow. When you go out to suppliers, EasyRFQ turns the same BOM into per-vendor RFQ packs, folds the returned quotes back in — selectively, quote by quote — and the bid tabulation scores the offers into a recommended mix.
Cost build-up by package and category, OPEX by component, engineering matrix, benchmarks, warnings and numbered assumptions — full audit trail.
Markdown · JSON · APIReal formulas, not pasted values: change margin, contingency, hours or strategy factors in the blue input cells and watch the estimate move.
XLSX · cent-exact vs engineAACE-style Word document: classification, methodology, versioned cost books, maintenance & I&C scenarios, exclusions and references.
DOCX · review-readyDesign office, estimator, project manager and buyer work on the same cost referential — role profiles scope what each one sees, and every project feeds the account’s cost capital.
Equipment list in, should-cost out in minutes: BOM take-off, factored bulks, the AACE range built into the result. The design-office profile works without vendor prices — safe to open wide.
Full AACE discipline — three data levels, every line explained and cited, versions diffed, parallel scenarios on one reference. The engine does the bookkeeping; you keep the judgement.
The AACE class determined from your deliverables — not declared — P10/P50/P90 with S-curve phasing, FTE staffing, scenario verdicts: the milestone file, ready for the review.
EasyRFQ supplier packs that never leak an internal price, multi-quote bid tabulation with a recommended mix, private observed-price memory, contractual index revision.
All plans run the same engine and the same rules — they differ by the simulation levels you unlock, the deliverables and the data privacy, never by rigour.
Costonomics is a small company, and you should plan for that. So the platform is engineered to make its vendor dispensable: your estimating capability must never depend on our fate — only your cost data updates do.
The engine is pure Python standard library — zero mandatory dependencies, no cloud runtime, no licence server, nothing phones home. On your infrastructure with the Platform licence, it keeps estimating whether we exist or not.
stdlib · on-premiseProjects, cost books, observed prices and reports live in plain JSON, CSV and markdown on your side — readable in any text editor, audit trail included. No export ransom: there is nothing to export, you already hold everything.
JSON · CSV · MarkdownSame input, same estimate, to the cent. The behavior is locked by 440+ automated tests you can run yourself, and the method is documented line by line — there is no proprietary magic to lose access to.
440+ tests · reproduciblePlatform licence option: the sources and each signed book release are deposited with every version. If we ever stop, the deposit is released and your team — or any Python developer — carries on. Continuity is contractual, not hoped for.
Platform licence optionIndependence is not a promise we make — it is how the product is built.