AstraMixAI
Concrete Mix Lab & Transparent Materials Market (100 Countries)

Design Concrete Mixes. Compare Real Material Costs.

AstraMix AI is organized into two focused tools for civil engineers, batching plants, and concrete buyers: Mix Lab for validated strength prediction and multi-objective mix optimization, and Materials Market for comparing concrete ingredients and ready-mixed concrete across verified local branches, service-area suppliers, and national buying channels.

AstraMix AI Concrete Mix Laboratory & Material Optimization
Section 1 · Mix LabSchema v1.0.0 · 1,030-Row Envelope

Strength Prediction & Multi-Objective Mix Optimization

Dedicated workspaces for predicting compressive strength from a known recipe or optimizing ingredient proportions under explicit constraints—with one-click transfer between both tools.

  • Allowed Input Ranges: Displays exact supported bounds (`kg/m³` and integer curing age `1–365` days) with sliders, steppers, and preset curing chips.
  • Three Understandable Result Cards: Predicted strength (MPa with RMSE uncertainty rules), Estimated ingredient cost (`/m³`), and Estimated embodied emissions (`kg CO₂e/m³`) with Good / Normal / Poor rating badges.
  • Separate w/c and w/b Ratios: Distinguishes `water / cement` from `water / (cement + slag + fly ash)` and never labels an infeasible solve as an optimum.
Section 2 · Materials Market100 Indexed Countries · Verified Supplier Directory

Ingredients, Dry Mixes & Ready-Mixed Concrete Directory

Send your Mix Lab recipe directly to the Materials Market or search standalone orders across 100 countries and 62,964 GeoNames cities with strict category separation and honest pricing provenance.

  • Country-Scoped GeoNames City Autocomplete: Search 62,964 cities across 100 countries (CC BY 4.0) with instant keyboard navigation and automatic reset on country change.
  • Progressive Discovery & Quantity Tiers: Separates verified local branches and outlets, service-area suppliers, and national buying channels, with automatic bulk quantity tier pricing and whole-bag rounding.
  • Strict Data Honesty: Separates branch-scoped stock from public price snapshots and quote-only suppliers. Customer reviews start empty (`0` fabricated ratings).
EMPIRICAL MODEL SCOPE

Supported Input Envelope vs. Engineering Qualification

The Mix Lab strength regressor (`XGBoost`, test RMSE `4.62 MPa`, MAE `3.03 MPa`, $R^2 = 0.917$) is bounded by the empirical 1,030-sample UCI Concrete Compressive Strength dataset.

Important distinction: These bounds define the statistical range where the machine learning model was trained. They are not a substitute for structural engineering codes (ACI 318, EN 206, GOST 27006) or laboratory trial batches.

Mix Lab Input Validation Contract (v1.0.0)

FeatureAllowed RangeUnitValidation Rule
Cement102.0 – 540.0kg/m³Required positive binder (`> 0` for `w/c`)
Blast Furnace Slag0.0 – 359.4kg/m³Included in total binder (`w/b`)
Fly Ash0.0 – 200.1kg/m³Included in total binder (`w/b`)
Mixing Water121.75 – 247.0kg/m³Full 1,030-row UCI minimum (`121.75 kg/m³`)
Superplasticizer0.0 – 32.2kg/m³High-range water reducer dosage
Coarse Aggregate801.0 – 1145.0kg/m³Crushed stone / gravel skeleton
Fine Aggregate594.0 – 992.6kg/m³Natural or manufactured concrete sand
Curing Age1 – 365 (integer)daysWhole integer days (`1, 3, 7, 14, 28, 56, 90, 180, 365`)