A web app for drafting CRediT (Contributor Roles Taxonomy) contribution statements for scholarly publications. Formerly released as CRediT Generator.
Add contributors, assign the 14 roles, and copy a manuscript-ready statement. The same grid is a contribution heatmap you can export as a figure, alongside files for journal submission systems: JATS4R XML, CSV, JSON, and Markdown.
CRediT Matrix is an independent project. It is not affiliated with or endorsed by NISO.
Try it: credit.duinlab.nl
- Contributors: add, rename, reorder, or paste a whole author list. Paste an ORCID iD or URL to look up the name, or a DOI to fill the whole list from the published record
- Ask your co-authors: send each person a link addressed to their own row. They tick what they did and send the link back; opening it fills in their row (roles, name, and iD) and nothing else
- Drafts: one per paper, switched from the header. They stay in this browser
- Contribution grid: select a cell to assign one of the 14 roles, as a yes/no value or as a contribution level. The grid is the heatmap, so you can transpose it, swap initials for full names, and recolor it
- Statements: group by role or by author, with full names or initials, and optional level labels. Copy as rich text or plain, and mark shared first authorship and corresponding authors
- Nine languages: the interface and the generated statement each pick their own language, so a Dutch interface can produce an English statement. Role names come from credit-translation; machine-readable exports keep the canonical English CRediT terms
- Exports: copy or download JATS4R XML, CSV, JSON, and Markdown, or download the heatmap as SVG or PNG
- Sharing and import: encode a draft in a URL, paste names, or import JSON, CSV, or JATS4R XML
- Works offline: install it, and a draft survives a flight. Only the ORCID and DOI lookups need a network
| First run | Statement & export |
|---|---|
 |
 |
TypeScript 6 throughout, on pnpm workspaces: the app at the root, reusable packages/core beside
it.
| Layer | Choice | Why |
|---|---|---|
| Frontend | Next.js 16 (App Router) | Runs on Cloudflare Workers via OpenNext |
| Styling | Tailwind CSS v4 | Design tokens via @theme; no runtime CSS |
| State | Zustand + immer + persist | Survives a refresh via localStorage |
| Validation | Zod | Schema checks at trust boundaries |
| Heatmap | Hand-crafted SVG (core) |
One SVG source feeds both download and canvas PNG |
| Offline | Service worker + manifest | Runtime cache of the app's own files; no build step |
Browser
└─ Next.js app (repo root, App Router)
├─ React UI + Zustand store (persisted to localStorage)
├─ @credit-generator/core ← all domain logic, runs in the browser
│ statements · JATS4R XML · CSV · JSON · Markdown · heatmap SVG · validation
└─ /api/orcid · /api/doi (route handlers) ──→ pub.orcid.org · api.crossref.org
← the only server-side calls
Nearly everything runs in the browser. packages/core holds the domain
logic as pure TypeScript, with zod as its only runtime dependency. XML import uses the native
DOMParser, and the PNG is drawn from the heatmap SVG onto a <canvas>.
The ORCID and DOI lookups are the exceptions, proxied by /api/orcid and /api/doi. ORCID's
public API sends no CORS headers. Crossref's polite-pool contact address belongs on the server,
not in every client bundle.
Contributions store a 0–100 integer score rather than a boolean, so the UI switches between
binary and level-based editing without changing the stored model. See
packages/core/README.md for the score-to-level boundaries.
No accounts, no server-side storage. This is a deliberate constraint, not a missing feature. A draft holds the names and ORCID iDs of co-authors who never visited this site. Keeping those in your browser means there is nothing to ask anyone to delete. Drafts are per-browser: move one between devices by exporting JSON. See ADR 0002 for what would have to change for this to be revisited.
Prerequisites: Node ≥ 26, pnpm ≥ 11, just (optional)
git clone https://github.com/simonvanlierde/credit-matrix
cd credit-matrix
pnpm install
pnpm dev # → http://localhost:3000CONTRIBUTING.md has the full command list and the lint/typecheck/test checklist.
Run just to list the watch/fix recipes layered on the pnpm scripts.
The live demo runs on Cloudflare Workers via
@opennextjs/cloudflare, which adapts the Next.js build. A
push to main builds and deploys it; CI lints, tests, and
build-checks, and never deploys.
To run the Worker yourself, set your own domain and bindings in wrangler.jsonc and open-next.config.ts:
pnpm preview # build + run the Worker locally
pnpm deploy # build + deploy to your Cloudflare accountBoth upstream proxies (/api/orcid, /api/doi) share one rate-limiter binding, API_RATE_LIMITER.
Without the binding, they fail open: the lookups keep working, unthrottled, and say so once in the
invocation log. After changing the binding, check the first deploy's logs for
API_RATE_LIMITER binding missing.
- Review the translations. Beside English, the interface ships in eight machine-translated
languages.
e2e/messages.spec.tschecks each catalog for key parity, balanced ICU braces, intact placeholders, and untranslated product and format names. No native speaker has reviewed the prose, though, and Japanese and Chinese need one most: the term for a contribution statement is unsettled. - Widen locale coverage. Eight translated locales ship today (de, es, fr, it, ja, nl, pt-PT,
zh-Hans), a curated subset of
credit-translation, vendored
under
packages/core/src/credit-i18n/translations. Refresh them withnode packages/core/scripts/fetch-credit-translations.mjs. - Read more from ORCID. The lookup takes the name only. Affiliation is the field submission systems ask for next.
- Make sharing discoverable. The onboarding barely hints at sharing a draft or asking a co-author to fill in their own row. Both are easy to miss entirely.
Bug reports and small features are welcome. See CONTRIBUTING.md for setup, testing, and the accessibility checks. PRODUCT.md states what the app is for and who it serves; DESIGN.md is the design system that UI changes are held to. Design decisions are recorded as ADRs.
CRediT Matrix builds on prior tools and scholarship on contributorship:
- The original Python/Dash CRediT Generator, which inspired this app.
- Role translations from credit-translation.
- The contribution matrix proposed by Nick Steinmetz (2019), which this app's heatmap descends from. Nature Index surveys it in "Researchers are embracing visual tools to give fair credit…".
- Brand, A., Allen, L., Altman, M., Hlava, M., & Scott, J. (2015). Beyond authorship: attribution, contribution, collaboration, and credit. Learned Publishing, 28(2), 151–155. https://doi.org/10.1002/leap.1210
- Holcombe, A. O., Kovács, M., Aust, F., & Aczel, B. (2020). Documenting contributions to scholarly articles using CRediT and tenzing. PLOS ONE, 15(12), e0244611. https://doi.org/10.1371/journal.pone.0244611
- Nakagawa, S., Ivimey-Cook, E. R., Grainger, M. J., O'Dea, R. E., et al. (2023). Method Reporting with Initials for Transparency (MeRIT) promotes more granularity and accountability for author contributions. Nature Communications, 14, 1788. https://doi.org/10.1038/s41467-023-37039-1
If you use CRediT Matrix in your work, please cite it. Metadata lives in CITATION.cff, and GitHub's "Cite this repository" button generates APA and BibTeX from it. The archived, versioned release is on Zenodo: doi:10.5281/zenodo.21213659.
van Lierde, S. CRediT Matrix [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.21213659
MIT © Simon van Lierde