Version: v0.1.0
Software type: Local-first research cockpit for traceable ESG disclosure evidence construction
Release status: Public SoftwareX reference artifact; immutable GitHub release v0.1.0
TRACE-ESG helps researchers register corporate sustainability reports, preserve source-linked candidate disclosure evidence, record human-review decisions, inspect descriptive summaries, and export reproducible evidence datasets. It is intended for provenance-aware research, teaching, demonstrations, and small local studies in which the connection from a source passage to a reviewed data row must remain inspectable.
TRACE-ESG v0.1.0 runs locally in a browser. No website is required, no backend is used, no database is used, and no public deployment is included.
TRACE-ESG constructs and manages disclosure evidence. It does not evaluate company ESG performance. Outputs are not ESG scores, ratings, rankings, compliance decisions, assurance findings, sustainability-performance assessments, or greenwashing classifications. Candidate evidence is provisional until a human reviewer records a decision.
All bundled demonstration companies, reports, claims, mock outputs, and validation records are fictional. They illustrate application behavior and file structure only. They are not empirical evidence and must not be used to make claims about real organizations.
- Local Body: the public v0.1.0 application, including the Research Cockpit, Source Vault, Evidence Ledger, Signal Workspace, Validation Workspace, descriptive dashboard, Research Export, local storage, and audit records.
- Controller Gateway: a controlled request/response boundary. In v0.1.0 it returns deterministic mock values for interface testing only.
- Protected Brain: not included in this release. No proprietary prompts, semantic classifiers, formulas, weighting systems, confidence calculations, scoring logic, rating logic, ranking logic, compliance logic, assurance logic, performance-assessment logic, or greenwashing-detection logic is published here.
- Source Vault: organizes fictional company and report metadata.
- Evidence Ledger: retains source-linked candidate evidence rows, locations, text, states, filters, and export controls.
- Validation Workspace: records accepted, rejected, needs-review, and not-reviewed decisions separately from source evidence.
- Research Export: produces CSV and JSON packages with software version, source identifiers, filters, row counts, validation context, timestamps, and claim-boundary metadata.
Mock Brain outputs displayed in the Signal Workspace are visibly labeled interface-testing placeholders. They are not analytical findings.
- Node.js 20.19 or a currently supported Node.js 22 release is recommended for full compatibility with the installed lint toolchain.
- npm 10 or later.
- A modern browser.
The release was build-verified on Windows with Node.js 20.12.2 and npm 10.5.0; a documented non-blocking oxlint engine warning applies to that older Node 20 point release.
From the repository root:
npm installnpm run devOpen the local URL printed by Vite. No server-side service or database setup is required.
npm run typechecknpm run buildThis creates a local dist/ folder. The folder is intentionally excluded from the clean source package. Building does not deploy the application.
trace-esg/
README.md
CITATION.cff
LICENSE.txt
package.json
package-lock.json
index.html
src/
public/
docs/
demo-data/
example-exports/
screenshots/
release-notes/
- Repository: https://github.com/chenla2026/trace-esg
- Immutable release: https://github.com/chenla2026/trace-esg/releases/tag/v0.1.0
- Default branch:
main - Annotated tag:
v0.1.0 - Tagged release commit:
e6215c4049320cde825e629d6b9dd064315f012a - License: MIT
No archive DOI has been created.
docs/ARCHITECTURE.mddocs/BODY_BRAIN_BOUNDARY.mddocs/CLAIM_BOUNDARY.mddocs/DATA_MODEL.mddocs/RESEARCH_WORKFLOW.mddocs/EXPORT_PACKAGE.mddocs/DEMO_DATA_NOTICE.mddocs/RELEASE_NOTES_V0_1_0.md
Version 0.1.0 uses browser localStorage, is single-user, and is designed for a local research cockpit rather than institutional-scale collaboration. It does not parse PDFs, perform OCR, fetch URLs, or extract semantic content automatically. It contains no real-company assessment dataset and no full research corpus. The software-build checks reported in the documentation do not constitute empirical validation of extraction accuracy, usability, or research findings.
Citation metadata is provided in CITATION.cff. Until the SoftwareX article and an archival identifier exist, cite the public software release as:
Sasaki, N., & Aung, E. D. (2026). TRACE-ESG (Version 0.1.0) [Computer software]. https://github.com/chenla2026/trace-esg/releases/tag/v0.1.0
TRACE-ESG v0.1.0 is released under the MIT License. See LICENSE.txt.
Nophea Sasaki: nophea.sasaki@sasin.edu (corresponding-author and interim software-support contact).