LWATI SCIENCE

A scientific workstation
in your browser.

Import data, run Python, analytical SQL and R, search scholarly evidence, connect claims to sources and preserve the path from question to result.

REAL COMPUTATION

The AI explains. The runtime calculates.

Numerical and statistical results should come from executed software. Lwati uses AI for assistance while deterministic runtimes do the work.

PY

Python

Browser-local CPython through Pyodide for data analysis and scientific workflows.

Local runtime
SQL

DuckDB

Query CSV, JSON and Parquet data with analytical SQL inside the browser.

Local runtime
R

webR

Run R and statistical workflows through WebAssembly without a dedicated R server.

Local runtime
EVIDENCE & PROVENANCE

Research with the evidence attached.

Search scholarly metadata, verify DOI-linked sources, create explicit claims, connect executed analyses to those claims and preserve dataset and code provenance.

QUESTION
↓
EVIDENCE
↓
DATASET
↓
ANALYSIS
↓
CLAIM
01

Local datasets

Scientific data has its own storage layer rather than being forced through small source-code payloads.

02

Reproducible results

Record runtime, code hash, dataset hash and execution information around analyses.

03

Evidence matrix

Keep sources, relationships and research claims inspectable instead of hiding them in generated prose.

04

Research Capsule

Export datasets, notebook context, evidence and provenance as a portable research artifact.

BRING A QUESTION

Run the calculation. Keep the trail.

Open Science ↗