TerraBrains is the operating system for geotechnical engineering. Capture boring logs offline on any tablet, run 213 code-checked analyses and 47 lab workflows, build a true 3D ground model with AI geological interpretation, and let AI draft your report — every claim cited. Sign deliverables that lock to a tamper-evident SHA-256 audit chain.
We are not replacing your CAD. We are replacing the spreadsheets, the email threads, and the “final-FINAL.pdf” chaos between them.
Borehole logs, SPT, samples — entered on a tablet that survives loss of signal, water, and gloves. Attach photos, voice notes, and scan sample barcodes. An offline outbox replays every record in order the moment you come back online.
Atterberg, sieves, triaxial, consolidation, and 43 more — each test queues into a real workflow with results that auto-classify and link straight back to the source sample.
AI drafts an engineer-grade report with supporting project or standard citations where available, and flags unsupported text for engineer review. Section diff, request-changes review loop, finalisation gate, then sign-and-stamp locks it to the immutable audit chain.
Where field-ops tools stop at a clean PDF, TerraBrains computes the engineering — 213 code-checked analyses, a true 3D ground model, and an AI that interprets it — all from the same boreholes, all citing their source.
Bearing capacity, settlement, piles (bored · driven · lateral · group), liquefaction, soil & rock slope stability, seepage, retaining & excavation systems, ground improvement, the full flexible + rigid pavement suite, tunnels and in-situ CPT / DMT / pressuremeter — to IS, Eurocode 7, NF P94-261, BS, AASHTO and ASTM. Not lookup tables: rigorous engines whose every result cites its governing clause, is covered by the golden test suite (190 cited validation suites document published-value benchmarks or formula-fidelity checks), and has adversarial tests for invalid and boundary inputs.
Beyond the number, design analyses can return a verdict — adequate, marginal, inadequate or over-conservative — with the governing criterion and a concrete next step. It reads on web, mobile and the AI assistant so reviewers can inspect the basis alongside the result.
Four premium suites attach to any plan: the Pavement Suite covers time-stepping Pavement-ME prediction (rutting, fatigue, thermal cracking, IRI and reliability), Burmister, IRC, AASHTO, MEPDG/JPCP distress models and FAARFIELD airfield design. The Advanced FE Suite provides in-house finite-element solvers including three-dimensional elasto-plastic continuum footing analysis, two-dimensional footing analysis, Richards/van Genuchten seepage, confined and unconfined seepage, strength-reduction slope, concrete slab and bearing checks. The Deep Foundations Suite covers axial t-z/q-w, lateral pile-group p-multipliers, the 6×6 pile-cap stiffness matrix and RC/steel pile sections. The Deep Excavation Suite covers embedded-wall soil-structure interaction with staged construction. Add or cancel any time.
Implicit RBF surfaces with structural-dip awareness, leave-one-out cross-validation, voxel volumetrics and smooth meshes — built straight from your borehole logs. Rotate the solid, read per-unit volumes, and see exactly where the model is well- or poorly-constrained.
It correlates units across boreholes, flags inconsistencies (outlier contacts, possible pinch-outs), weighs the model's cross-validation error, and drafts a report-ready geological narrative. Advisory by design — you accept, edit, or reject every line, and nothing is auto-signed.
Import legacy logs — including AI-assisted ingestion of scanned and PDF bore logs — and export validated AGS 4 and DIGGS 2.6 for common FE packages and your peers. Open standards in, open standards out: your data is never locked in.
Every paid plan includes 165 base analyses and 47 lab tests; the remaining 48 optional specialist analyses are in Store packs. Every catalogue item carries its governing standard and a worked example you can open.
Legacy per-seat tools often cost $3,000–$5,000/seat/yr. We're positioned at the AI-native, audit-grade end of that range — and we're honest about it.
If you don’t see your question here, ping hello@terrabrains.ai.