treo.io
■ treo.io process engineering engineBuilt by the co-founder of METSIM InternationalConcept to PFS

The engine that learns from every model it builds.

treo.io is an AI engine trained on decades of METSIM® models, our audit history and published metallurgy. Give it a resource report, or a photo of a sample, and it returns a solved mass and energy balance and a full PFD set in hours, ready for specialist review.

A live PFD sheet, generated from a solved METSIM® model. Lines, stream numbers and the stream table all come from the balance.

Issued
1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 A A B B C C D D E E F F G G H H REV DATE DESCRIPTION BY CHK APP A 25.10.26 ISSUED FOR INFORMATION (BFD / PFD SET) AI - - B 31.10.26 RE-DRAWN TO TREO.IO PFD STANDARD REV 0 AI AI QC - treo.io PROCESS ENGINEERING NOT FOR CONSTRUCTION KOLWEZI Cu-Co OXIDE PROJECT - PRE-FEASIBILITY STUDY (TREO.IO DEMONSTRATION) BALANCE: METSIM® MODEL REV_0009 (CONVERGED REV_0004 BASIS) - DESIGN THROUGHPUT 3.0 Mt/a, 372 t/h DRY ORE PROCESS FLOW DIAGRAM AREA 2200 REDUCTIVE LEACH DRAWING No. KCC-PF-2200-001 SHEET 5 OF 17 SCALE / SIZE NTS / A1 REV B M M M M M M 2200-TK-001 2200-AG-001 LEACH TANK 1 2200-TK-002 2200-AG-002 LEACH TANK 2 2200-TK-003 2200-AG-003 LEACH TANK 3 2200-TK-004 2200-AG-004 LEACH TANK 4 2200-TK-005 2200-AG-005 LEACH TANK 5 KCC-PF-3100-001 FROM 3100-PU-002A/B HG RAFFINATE TO LEACH 092 KCC-PF-4200-001 FROM 4200-PU-001A/B Cu RESIDUE RECYCLE TO LEACH 153 KCC-PF-4400-001 FROM 4400-PU-001A/B Co SCAVENGER RESIDUE RECYCLE TO LEACH 182 KCC-PF-6100-001 FROM 6100-PU-001A/B SULPHURIC ACID TO LEACH 401 KCC-PF-2100-001 FROM 2100-PU-001A/B PRE-LEACH THICKENER UNDERFLOW TO LEACH 020 KCC-PF-6100-001 FROM 6100-PK-001 SO2 GAS TO LEACH TANK 1 411 KCC-PF-6100-001 FROM 6100-PK-001 SO2 GAS TO LEACH TANK 2 (STANDBY) 412 030 031 032 033 KCC-PF-2300-001 TO 2300-TH-001 LEACH DISCHARGE TO LEACH THICKENER 034 KCC-PF-6100-001 TO 6100-PK-006 LEACH VENT GAS TO SO2 SCRUBBER 520 VENT HEADER BUNDED AREA - LEACH TANKS 2200-SH-001 SAFETY SHOWER AND EYEWASH POTABLE WATER 2200-SP-001 LEACH AREA BUND SUMP AND PUMP TO 2200-TK-001 FLOOR SPILLAGE STREAM 020 TOTAL 715.9 t/h SOLIDS 52.0 % w/w FLOW 473.1 m³/h STREAM 092 TOTAL 376.6 t/h Cu 1.01 g/L H2SO4 29.6 g/L STREAM 401 TOTAL 32.1 t/h H2SO4 1,812.4 g/L FLOW 17.5 m³/h STREAM 411 GAS 1.8 t/h GAS 647 Nm³/h STREAM 034 TOTAL 1,120.8 t/h SOLIDS 31.6 % w/w Cu 12.56 g/L STREAM 520 GAS 12.1 t/h GAS 6,119 Nm³/h EQUIPMENT LIST TAG DESCRIPTION D / S STATUS SIZE / kW 2200-TK-001 LEACH TANK 1 1D NEW 1,163 m³, 160 kW 2200-AG-001 AGITATOR, LEACH TANK 1 1D NEW 2200-TK-002 LEACH TANK 2 1D NEW 1,163 m³, 160 kW 2200-AG-002 AGITATOR, LEACH TANK 2 1D NEW 2200-TK-003 LEACH TANK 3 1D NEW 1,163 m³, 160 kW 2200-AG-003 AGITATOR, LEACH TANK 3 1D NEW 2200-TK-004 LEACH TANK 4 1D NEW 1,163 m³, 160 kW 2200-AG-004 AGITATOR, LEACH TANK 4 1D NEW 2200-TK-005 LEACH TANK 5 1D NEW 1,163 m³, 160 kW 2200-AG-005 AGITATOR, LEACH TANK 5 1D NEW 2200-SH-001 SAFETY SHOWER AND EYEWASH 1D NEW* 2200-SP-001 LEACH AREA BUND SUMP AND PUMP 1D NEW* NOTES 1. MASS BALANCE: METSIM® MODEL KOLWEZI CuCo PFS REV_0009 (CONVERGED, DESIGN CASE 3.0 Mt/a AT 92 % AVAILABILITY, 8,059 h/a). FLOWS ARE NOMINAL DESIGN t/h. 2. VOLUMETRIC FLOWS AND SGs ARE ESTIMATED FROM COMPONENT DENSITIES (SOLIDS), A SULPHATE- SOLUTION CORRELATION (AQUEOUS) AND ORGANIC SG 0.82. MODEL STREAM TEMPERATURES ARE NOMINAL; THE HEAT BALANCE IS NOT MODELLED. 3. MODEL STREAM REF. IN THE STREAM TABLE GIVES THE METSIM® STREAM NUMBER(S) EACH TREO.IO STREAM CARRIES. 4. LEACH TANKS ARE COVERED AND VENTED TO THE SO2 SCRUBBER 6100-PK-006 VIA A COMMON VENT HEADER (STREAM 520 = SUM OF THE FIVE TANK VENTS). 5. SO2 IS SPARGED BELOW THE AGITATOR OF TANK 1; THE TANK 2 SPARGER (STREAM 412) IS A STANDBY ADDITION POINT AND CARRIES NO FLOW IN THE DESIGN CASE. 6. LEACH AREA IS BUNDED; BUND CAPACITY 110 % OF THE LARGEST TANK. 7. Cu AND Co SCAVENGER RESIDUES ARE RECYCLED TO TANK 1 TO RE-LEACH CO-PRECIPITATED Cu AND Co. STREAM No. UNITS 020 030 031 032 033 034 092 153 182 401 411 412 520 DESCRIPTION PRE-LEACH THICKENER LEACH TANK 1 TO LEACH TANK 2 TO LEACH TANK 3 TO LEACH TANK 4 TO LEACH DISCHARGE TO HG RAFFINATE TO Cu RESIDUE RECYCLE Co SCAVENGER SULPHURIC ACID TO SO2 GAS TO LEACH SO2 GAS TO LEACH LEACH VENT GAS TO UNDERFLOW TO LEACH TANK 2 TANK 3 TANK 4 TANK 5 LEACH THICKENER LEACH TO LEACH RESIDUE RECYCLE TO LEACH TANK 1 TANK 2 (STANDBY) SO2 SCRUBBER LEACH SOLIDS t/h 372.3 354.4 354.4 354.4 354.4 354.4 - 0.6 1.0 - - - - LIQUID (AQ + ORG) t/h 343.6 766.4 766.4 766.4 766.4 766.4 376.6 1.9 3.0 32.1 - - - GAS t/h - - - - - - - - - - 1.8 - 12.1 TOTAL t/h 715.9 1,120.8 1,120.8 1,120.8 1,120.8 1,120.8 376.6 2.5 3.9 32.1 1.8 <0.1 12.1 SLURRY / LIQUID FLOW m³/h 473.1 828.1 828.1 828.1 828.1 828.1 347.1 2.1 3.2 17.5 - - - GAS FLOW Nm³/h - - - - - - - - - - 647 - 6,119 SOLIDS SG - 2.77 2.70 2.70 2.70 2.70 2.70 - 2.52 2.59 - - - - LIQUID SG - 1.014 1.100 1.100 1.100 1.100 1.100 1.085 1.039 1.036 1.840 - - - SLURRY SG - 1.513 1.353 1.353 1.353 1.353 1.353 - 1.218 1.219 - - - - SOLIDS % w/w 52.0 31.6 31.6 31.6 31.6 31.6 - 25.0 25.0 - - - - P80 µm - - - - - - - - - - - - - Cu IN SOLIDS % w/w 2.50 0.27 0.27 0.27 0.27 0.27 - 9.87 0.01 - - - - Co IN SOLIDS % w/w 0.50 0.11 0.11 0.11 0.11 0.11 - 6.82 10.62 - - - - Cu IN SOLUTION g/L - 12.56 12.56 12.56 12.56 12.56 1.01 - - - - - - Co IN SOLUTION g/L - 4.59 4.59 4.59 4.59 4.59 4.48 2.33 0.01 - - - - H2SO4 IN SOLUTION g/L - 8.0 8.0 8.0 8.0 8.0 29.6 - - 1,812.4 - - - Cu IN ORGANIC g/L - - - - - - - - - - - - - TEMPERATURE °C 20 20 20 20 20 20 20 20 20 20 20 - 20 MODEL STREAM REF. 2103 (=2102) 2206 2207 2208 2209 2210 2201 2204 (=4205) 2205 (=4404) 2202 2203 2211 2212-2216 YOUR LOGO SAMPLE CLIENT · SAMPLE PROJECT · PRE-FEASIBILITY STUDY PROCESS FLOW DIAGRAM AREA 2200 REDUCTIVE LEACH DRAWING No. TRE-2200-PFD-001 SHEET5 OF 17 SIZEA1 REVB
1 Process and utility lines drawn from the model's stream connectivity 2 Stream numbers placed and checked against all model streams 3 Stream table filled from the solved balance 4 Title block, symbols and numbering in your standard
You provide
Anything from a sample photo and location to a resource report, mineralogy and testwork
We start from
A proven METSIM® model for your commodity and process route
You receive
Audited mass and energy balance, PFDs with stream tables, assumptions register
Time to first package
Hours, not weeks. Specialist-reviewed issue in days
Calvin MorrisonCo-founder, METSIM InternationalFounder, treo.io METSIM® is the industry-standard metallurgical process simulator.
METSIM® models built and audited for
  • SENET
  • DRA
  • Process Ideas
Time to first package

An engineering team takes weeks. The AI engine takes hours.

Same scope: flowsheet, solved mass and energy balance, audit and a full PFD package with stream tables.

12–16weeks, typical conventional delivery
Hoursfor the AI engine to produce the first package
Daysto specialist review and issue

Conventional durations are indicative for a concept to PFS package and vary with flowsheet complexity. Specialist review and testwork follow on both paths.

1.0AutoCAD deliverables

Native AutoCAD drawings your drafters can pick up and run with.

Every PFD is issued as an AutoCAD .dwg on your CAD standard, with the balance carried inside the drawing as block attributes. Nothing is redrawn at the handover to detailed design, and no stream data is keyed in twice.

KCC-PF-2200-001.dwgCommand: DATAEXTRACTION
Layer properties
NameOnLinetype
PFD-PROCESS●Continuous
PFD-SECONDARY●Continuous
PFD-STANDBY●DASHED
PFD-EQUIPMENT●Continuous
PFD-STREAM-ID●Continuous
PFD-STREAM-DATA●Continuous
PFD-EQUIP-TAG●Continuous
MODEL-ONLY○Continuous
TITLE●Continuous

Layer names, colours and linetypes follow your template (.dwt).

Properties · Block reference STREAM_ID

Selected block cycles through the sheet. Values are written from the solved model.

D-01

Your CAD standard

Layers, linetypes, text styles, blocks and title block from your own template, so the drawings drop straight into your drawing register.

D-02

Attributed blocks

Equipment tags and stream numbers are blocks carrying tag, description, duty and stream data, not exploded text.

D-03

Data extraction ready

Run DATAEXTRACTION and the equipment list and stream table come straight out of the drawing into Excel.

D-04

Model-linked revisions

When the model changes, drawings are reissued with updated attributes, the revision block filled in and the changes listed.

D-05

Registers included

Drawing register, stream register and equipment list issued with every package, tied to the same tags.

D-06

Straight into P&IDs

Your P&ID and detailed design teams start from the same tags, line data and design conditions, which cuts the handover from weeks to days.

PDF sets are issued alongside for review and sign-off.

2.0Why treo.io

AI-led. Proven routes. Testwork where it counts.

Most deposits can be treated by a process route that already works somewhere. treo.io starts from that route, puts every gap on the record, and points the testwork at the gaps that matter. treo.io does not offer packages as a substitute for testwork.

2.1

Proven routes, not first principles

Each package starts from a METSIM® model of a route that already operates, for your commodity and ore type. The engine adapts it to your resource and mineralogy.

2.2

Testwork where it changes the answer

Every assumption is run across its known range. The ones that move recovery, reagent use or equipment size most become the first testwork priorities.

2.3

Every assumption on the record

Each value is tagged as basis, client data or a numbered assumption, with its source, its range and the action that will close it.

2.4

One model, no retyping

Stream tables, drawings, the assumptions register and the balance workbook are generated from the same solved model, so there is no manual transcription between them.

2.5

Reviewed by a specialist in your metal

An independent industry reviewer for your commodity checks the route selection, the assumptions and the results before issue.

2.6

Precision as the data arrives

Testwork results replace assumptions in the same model. The package is reissued with the change log, ready for detailed design.

Conventional early-stage studytreo.io
Who does the workProcess engineers and drafters, by handAI engine builds; engineers review and sign off
Starting pointBlank model, built by whoever is availableProven model for the commodity and route
Inputs to startA developed flowsheet and testworkA sample photo and location, upwards; public data fills the gaps
TestworkScoped before the flowsheet is understoodTargeted at the assumptions with the largest impact
AssumptionsOften implicit or buried in spreadsheetsStated, referenced and linked to a validation action
ReviewInternal, generalistIndependent commodity specialist
Drawings and tablesDrafted separately and reconciled by handGenerated from the solved model, with no retyping
Time to first package12–16 weeks, for a team of engineersHours, for the AI engine
3.0Track record

Decades of models built. Now AI built, specialist reviewed.

The engine is built on the models our team has developed, audited and run over decades, across operating plants and studies from concept to feasibility. Select a part of the METSIM® emblem to filter the list by discipline.

Looking backDecades of models builtEngineered by hand across operating plants and studies, then audited and kept.
TodayAI built, specialist reviewedThe engine builds from that library in hours. Industry specialists sign off.
Looking forwardEvery model adds to the libraryEach solved study and scenario becomes a reference for the next.
  • Hydrometallurgy
  • Ore and beneficiation
  • Metals and refining
  • Plant and utilities

METSIM® and the METSIM emblem are trademarks of Metsim International.

CommodityProcessScopeRegion

4.0How the engine learns

Model upon model, it gets better.

The engine isn't a general-purpose chatbot pointed at METSIM®. It works from a curated body of process knowledge: our own model library, published metallurgy, and the findings of every audit and review it has been through. Each project adds to it.

Layer 1

Our model library

Decades of METSIM® models: process route structures, unit-operation set-ups, controller strategies and solver settings that have already converged on real flowsheets.

Proprietary to treo.io
Layer 2

Published metallurgy

Conference papers, filed technical reports and operating data. Every extracted value is stored with its source and a verbatim quote, so any number can be traced back to where it came from.

Public domain, curated
Layer 3

Your project data

Your mineralogy, testwork and drawing standards. Used on your projects only and never shown to another client. Only anonymised factors, such as a leach extent range, return to the shared library, under the terms of your agreement.

Yours, kept separate
Libraries routesrangesrules +1

Every finding becomes a rule

Each package passes through an independent check, a structured audit and a commodity specialist. Anything they catch is written back as a rule, a check or a tighter parameter range, and applied on every job after that.

When testwork arrives, measured values replace assumptions in your model, and the anonymised range they fall in updates the library for the next study on similar ore.

The engine improves through its libraries, checks and rules. It is not retrained on your data.

Self-built reference library

Every flowsheet it solves makes the next one faster

For each study the engine solves a family of models, not one: the base case plus throughput, grade, ore-type and process-route alternatives. Each scenario is solved, closure-checked and stored with its inputs, assumptions and results.

Model upon model, this builds a library of solved flowsheets far larger than any team could assemble by hand. A new study starts from the nearest solved case for that commodity and ore, not from a blank template.

Engine-built scenarios are tagged as model results. They rank below plant and testwork data, and are recalibrated when measured data arrives.

Base caseThroughputGrade and ore typeRoute alternative

Illustrative: solved scenarios added per study

SourceWhat was foundWhat the engine does now
Cu oxide, ZambiaIndependent check found a solute purge hidden behind a fixed flow to the tailings damTests every fixed flow for what it may be masking before issue
Cu oxide, ZambiaAudit found acid dilution heat missing and an implausible heat input at the tailings damRuns a plausibility check on every heat term, not only overall closure
Cu oxide, ZambiaA model was issued before its hygiene audit was completeAudit and hygiene are part of the definition of done; nothing ships without them
Cu-Co, DRCFresh acid consumption came out 25–64 % above published district valuesBenchmarks reagent use against the published range for the district before issue
Cu-Co, DRCRecycle convergence stalled on near-zero trace flowsSets solver tolerances for trace components when the model is built
Cu-Co, DRC148 model streams had to be reconciled with the drawingsAn automated drawing checker accounts for every model stream on every sheet
5.0Inputs

Start with what you have. Even a photo of a sample.

You don't need a resource report to start. The engine fills the gaps from public data: filed technical reports, conference papers, regional mineralogy and operating analogues. Every value it uses is tagged with its source. Each input you add replaces assumptions and tightens the accuracy class.

L0

Sample photo and location

Ore type, likely mineralogy and district analogues inferred and stated as assumptions.

Concept · AACE Class 5
L1

Public resource data

Grades, tonnages and deportment from filed JORC, NI 43-101, SAMREC or S-K 1300 reports.

Scoping · Class 5
L2

Your mineralogy

XRD, bulk mineral analysis and QEMSCAN replace inferred phases and deportment.

Scoping to PFS · Class 4–5
L3

Your testwork

Leach, flotation, thickening and SX results replace the assumptions ranked highest by impact.

PFS · Class 4
Public-domain sources the engine reads
Filed technical reportsSEC S-K 1300, SEDAR+ NI 43-101, ASX JORC releases, JSE SAMREC
Conference proceedingsSAIMM, AusIMM, CIM and peer-reviewed journals
Operating analoguesPublished design and operating data from plants on similar ore
Geological surveysRegional mineralogy, deposit models and assay databases
731metallurgical values extracted from 41 conference papers for one Cu-Co study, each stored with a verbatim source quote

What public data can and can't do. It is usually enough to choose a process route with a good chance of holding up at PFS and to produce a PFS-format package at concept or scoping accuracy. Reaching PFS accuracy needs testwork on your ore. The engine tells you which tests to run first.

6.0Commodity library

A proven model for your metal.

Each commodity has its own library of METSIM® process-route models, tested against operating data and completed studies, with a specialist reviewer assigned. Select a commodity to see its routes.

Proven routes in libraryRoutes in validation
7.0Engine logic

How a package is calculated.

The AI engine follows the same chain a metallurgist does, in minutes per step rather than days: mineralogy sets the chemistry, the chemistry sets the balance, and the balance shows which assumptions need testwork first. The example below is a copper oxide leach.

Mineralogy sets the chemistry

Copper deportment by phase, from XRD, bulk mineral analysis and QEMSCAN, decides which reactions the model needs and how much copper each one can release.

Cui = di · Cufeed    Σ di = 1

Source: client mineralogy report

PhaseFormuladi, % of Cu
MalachiteCu₂CO₃(OH)₂46
ChrysocollaCuSiO₃·2H₂O28
CupriteCu₂O9
ChalcopyriteCuFeS₂12
Other–5

Worked example with illustrative values. Every number in a delivered package traces to the design basis, the client's data or a numbered assumption.

8.0Ask the model

A METSIM® model you can talk to.

Ask your model a question in plain language. The engine reads the solved balance, runs the case if it needs to, and answers with the numbers and where they came from. No need to open METSIM® or wait for the modeller.

Kashime-type Cu oxide · model Rev A1Early access
Engine activity
    You can ask
    • What if throughput goes up 20 %?
    • Why is lime consumption so high?
    • Which assumption matters most?
    • Show me stream 9.

    Example conversation with illustrative values.

    9.0Reference libraries

    The data behind every model.

    The models are only as good as the data under them. Each package draws on curated reference libraries that grow with every project.

    RL-01

    Process route templates

    Proven flowsheet structures per commodity and ore type, with unit operations, recycles and control philosophy.

    RL-02

    Chemistry and thermodynamics

    Reaction sets, mineral phases, extents and heat data, matched to the mineralogy you report.

    RL-03

    Mineralogy to recovery

    Deportment-based recovery and reagent relationships, from testwork and plant records.

    RL-04

    Reagent and utility benchmarks

    Consumption ranges for acid, lime, cyanide, flocculant, power, water and steam by route.

    RL-05

    Equipment design factors

    Residence times, loadings, unit areas and efficiencies for sizing at concept and PFS level.

    RL-06

    Drawing standards

    Symbol libraries, title blocks and numbering conventions, held per client and applied to every drawing.

    10.0Method

    From resource statement to PFS package.

    Each step has an owner. The loop closes when testwork replaces the last material assumption.

    1. 10.1

      Resource and mineralogy

      Grades, tonnages and ore types from the resource report; phases and deportment from mineralogy.

      Your team
    2. 10.2

      Route selection

      The best-matched proven route is selected from the commodity library and agreed with you.

      AI engine + reviewer
    3. 10.3

      Design basis

      Basis reconciled up front. Gaps become stated assumptions with validation actions.

      AI engine
    4. 10.4

      Model build

      The AI engine adapts the route model to your ore, grades and throughput, and solves it. Minutes to hours, not weeks.

      AI engine
    5. 10.5

      Audit

      Checklist audit, elemental, water and heat closure, and an engineering sense-check.

      AI engine
    6. 10.6

      PFD package

      The AI engine generates the drawings and stream tables from the audited model, on your standards.

      AI engine
    7. 10.7

      Specialist review and issue

      The commodity reviewer signs off. The package is issued with the assumptions register and a ranked testwork programme.

      Reviewer
    8. 10.8

      Testwork and reissue

      Results replace assumptions in the model, and the package is reissued at the precision the next study stage needs.

      Your team + treo.io

    Built to feed your technical report. The balance, recoveries, reagent consumptions and assumptions register are structured to support the metallurgy and recovery-method content expected under JORC, NI 43-101 (Items 13 and 17) and SAMREC. Responsibility for the report stays with your Competent or Qualified Person.

    11.0Reviewer panel

    Reviewed by specialists who use it on their own projects.

    Every package is signed off by an independent industry specialist. Our reviewers aren't only checking the engine's work: they run treo.io on their own studies, and what they find on those jobs feeds back into the rules and libraries.

    ReviewersSign off route selection, assumptions and results before any package is issued
    UsersRun the engine on their own consulting and study work
    ContributorsTheir corrections become rules and tighter ranges for every client
    12.0Products

    What you can order.

    TB-01

    Treo BalancePowered by METSIM®

    Steady-state mass and energy balance of the whole flowsheet, with reagents, utilities and power. Delivered with the native model file, the balance workbook, the assumptions register and the audit.

    Available
    TP-02

    Treo PFDDrawn from the balance

    The full process flow diagram set with stream tables tied to the model, on your symbols, title blocks and numbering.

    Available
    TA-04

    Treo AskConverse with your model

    Ask questions of the delivered model in plain language: what-if cases, sources of any number, sensitivities and stream data, answered from the solved balance.

    Early access
    TI-03

    Treo P&IDFrom PFD to instrumented design

    Line lists, valves and control loops carried forward from the PFDs and the balance's control philosophy.

    In development
    13.0Case study

    Two copper projects.

    One started from a client's design basis. The other started from a photograph of a sample and public data.

    Client project · Zambia

    Copper oxide leach, SX and EW, from a design basis

    Flowsheet
    Dual HG/LG leach with SO₂, six-stage CCD, SX, EW, neutralisation and tailings
    Study level
    Concept to pre-feasibility
    Inputs
    Design basis, block flow diagram, mineralogy and Cu deportment
    Delivered
    Audited model, balance, assumptions register, PFD pack
    The balance surfaced design issues, including PLS cooling and a manganese bleed, and sent them back to the client's team before PFS.
    20ktpaCopper cathode from a 1.86 Mt/a feed
    0.02%Elemental closure, full flowsheet
    80%Overall copper recovery on stated assumptions
    6Basis and design questions raised before PFS
    Internal demonstration · DRC

    Copper-cobalt oxide, from a sample photo

    Inputs
    One photograph of a Kolwezi-district sample. No resource report, no client data.
    Engine
    Inferred ore type and mineralogy, harvested public metallurgy data, selected a proven agitated leach, CCD, SX-EW and cobalt hydroxide route
    Delivered
    Converged METSIM® model, 17 A1 PFD sheets, equipment list, Class 5 cost model, validation register
    Benchmarking against public data flagged the model's own acid consumption as 25–64 % above district references, and set it as the first item to recalibrate.
    65.8kt/aCopper cathode at 3.0 Mt/a, 87.7 % recovery
    10.2kt/aCobalt in hydroxide, 68.2 % recovery
    148Model streams, all accounted for on the PFDs
    25Items on the validation register, each with a source
    14.0Who it's for

    Inside your team, under your name.

    treo.io isn't an engineering house competing for your work. We deliver under your name or ours.

    EPCs and study firms

    • White-label delivery on your standards, to your client.
    • We never bid against you on work you bring us.
    • Your data stays yours. Project data is never shared with other clients.
    • Framework terms with priority turnaround and fixed per-package rates.

    Owners and developers

    • Test the concept from a resource report before committing to a full study.
    • Compare routes and throughputs on the same basis.
    • Brief your EPC with a basis, a balance and a list of known gaps.
    • Keep the model current as testwork lands.
    15.0Book a demo

    Your next study starts where the last hundred left off.

    Bring a resource report, or a photo of a sample. In 30 minutes we'll walk through a complete package for a commodity close to yours, from the inputs to the issued drawings.

    Prefer to pick a time? Book a 30-minute demo →

    We'll use these details only to arrange the demo.