NEWLattice Graph 4.0 — multimodal materials search

Turn decades of materials data into decisions your engineers can ship.

Lattice Graph ingests your lab notebooks, characterization files, and supplier specs — and returns the one composition that meets your constraints. No more searching through PDFs.

Indexed assays
2.4M
Median search
340ms
Formats
40+
42 nodes · stable
sample AL-7075-T6
σy = 503 MPa
cluster coherence 0.947
Works with the formats your lab already produces
XRD
FTIR
DSC
Tensile CSV
Lab PDFs
Supplier Specs
Impact

Built for teams who ship atoms, not just apps.

What our platform is designed to deliver — from aerospace alloys to battery cathodes to next-generation ceramics.

Cycle time
10×
Faster from data request to candidate shortlist, based on internal benchmarks.
First-try yield
68%
Of recommended compositions meet target specs on the first physical run.
Time to insight
1 day
Median onboarding from dataset upload to first production-ready query.
Data indexed
2.4M assays
Across 40+ internal formats — XRD, FTIR, tensile, DSC, and lab notebook PDFs.
How it works

Three steps from fragmented data to engineering-grade answers.

No data-science team required. Lattice learns your naming conventions, infers missing metadata, and surfaces provenance for every recommendation.

01CONNECT

Ingest messy data as-is

Drop in 40+ formats — lab notebook PDFs, instrument exports, supplier CSVs. No schema mapping required.

02STRUCTURE

Auto-unify into a graph

Lattice links compositions, treatments, and outcomes into a living knowledge graph — with every edge traceable to source.

03DECIDE

Query in plain language

Ask "which Mg alloys survive 400°C creep at under $9/kg?" — get a ranked shortlist with confidence intervals.

Use cases

Problems we hear about in every first call.

Lattice is purpose-built for the decisions materials engineers actually make in the week before a design review.

Alloy selection

Shortlist candidate alloys against multi-objective constraints.

Trade off yield strength, density, supply risk, and unit cost on the same plot. Pareto fronts update live.

Failure analysis

Find the closest prior failure in seconds.

Paste a micrograph or spec — Lattice returns similar prior cases from your own archive and the open literature.

Supplier risk

Dual-source before procurement asks.

Automated spec matching across 28 supplier catalogs.

Regulatory

RoHS & REACH sanity checks.

Flag restricted substances before BOM lock.

Process

Heat-treat optimization.

Recommend cycles from prior runs that hit your spec window.

Composition search

Inverse design with targeted properties.

Specify the property profile you need — Lattice proposes compositions with uncertainty bounds grounded in your own data.

Handoff

Export to your existing PLM and ELN.

One-click sync to Teamcenter, Windchill, LabArchives, and whatever CSV lives in your shared drive.

Inside Lattice

Familiar tools, purpose-built for materials.

Built for engineers, not data scientists. Use it in the browser, or pipe it into your existing notebooks via the SDK.

app.latticelabs.dev / projects / airframe-2027 / alloy-search
Workspace
Overview
Datasets214
Alloy search12
Insights
Filters
Aluminum families48
Titanium alloys22
Steels (stainless)31
Composites17
Strength / density trade-off
ScatterParetoTable
Candidates
48↑ 12
On Pareto
7
Match score
0.921↑ 0.04
Selected · AL-7075-T6
density2.81 g/cm³
yield_str503 MPa
elongation11 %
cost_idx$8.40/kg
fatigue_10⁷159 MPa
corrosionCl⁻ mild
suppliers14
confidence0.947
n_samples1,208
Developer SDK

One line from notebook to knowledge graph.

A batteries-included Python SDK with typed responses, streaming results, and optimistic caching. Use it in a Jupyter cell the same way you'd query a database.

query.pyresponse.jsonpython 3.11
from lattice import Client, spec

client = Client(api_key="sk_live_…")

# natural-language composition search
results = client.search(
    "Mg alloys stable above 400°C "
    "with yield strength > 220 MPa "
    "and cost below $9/kg",
    filter=spec.source(tier="internal"),
    limit=12,
)

for m in results.shortlist:
    print(m.composition, m.confidence)

# → WE43-T6       0.94
# → ZK60A-T5      0.91
# → Elektron 21   0.87
Early access
WHY WE'RE BUILDING THIS
Materials teams still rely on shared drives, lab notebooks, and aging SharePoint sites that nobody fully trusts. We think they deserve better.

We're building the platform we wished existed when we were materials engineers ourselves.

The Lattice Graph Team
Currently in early access

Ship the material, not the search for it.

We're onboarding early teams now. Sign up for early access and help shape the platform.