CorrespExplorer is a browser-based research tool for exploring correspondence metadata in CMIF and TEI letter records. Researchers can examine correspondence networks, places, dates and topics through linked views, filter a shared selection and trace patterns back to individual letter records and their editions.
Open the live demo to explore the Hugo Schuchardt Archiv, or load your own files. Local files are processed in your browser, without installation or server-side data processing.
- Open the live demo with the bundled Hugo Schuchardt Archiv, or choose “Eigene Daten” to load local files or a source URL.
- Select “Im Explorer öffnen” and explore letters, persons, places, time and topics. The interface is in German.
- Filter the shared selection and open its letter records to inspect the evidence behind a pattern. Follow the links to the source editions.
You can compare correspondents, collect records in the Wissenskorb and export selections as CSV or JSON. Shared filter URLs work with datasets that recipients can also access. They do not upload or share your local files.
CorrespExplorer reads CMIF, the TEI-based Correspondence Metadata Interchange Format, and TEI letter records containing correspDesc. Several files from one edition can be loaded as a single dataset. You can also load results from correspSearch.
What each view can show depends on the metadata supplied by the edition. The parser preserves supported editorial qualifications, postmarks, addresses and other letter-level information. Data documentation describes the supported structures and datasets, and the uncertainty specification explains how missing and uncertain values are handled.
Source URLs, correspSearch queries and optional Wikidata enrichment contact external services. The OpenFreeMap basemap is requested when the map view is first opened. Libraries and fonts are bundled locally.
Version 0.9.0 is available on GitHub Pages. The v0.9.0 pre-release provides the tagged source. CorrespExplorer is a research prototype without a support commitment, and findings should be checked against the underlying editions. It displays metadata and links to editions, without displaying letter texts or editing source data.
Module tests, browser smoke tests and an independent agent review cover the documented preview. Scholarly validation, manual tasks and accessibility checks remain incomplete. Tests are run locally and are not part of continuous integration. The changelog records the current verification status and known limitations, with detailed evidence in testing.md and outstanding work in plan.md.
Serve the website locally with Python:
python -m http.server 8000 --directory docsOpen localhost:8000. Run the module tests at test.html?test=true, or run the browser smoke test with uv run tests/browser/smoke.py. The testing guide includes the fixed task set and further checks.
The application uses JavaScript ES modules, D3, MapLibre GL JS and noUiSlider, without a build step. GitHub Pages publishes docs/ from main. Dataset preparation scripts are in preprocessing/.
Development follows Promptotyping, a working practice that translates research data and their contexts into research artefacts with LLM agents. The project knowledge documents scope, data, design and decisions, mostly in German. The journal records the development history. Coding instructions are in CLAUDE.md.
Code is licensed under MIT, documentation under CC BY 4.0. Third-party datasets, libraries, fonts and marks retain their own terms.
The bundled Hugo Schuchardt Archiv export is published by the Centre for Information Modelling, University of Graz, under CC BY 4.0. Its JSON derivatives are converted, indexed and enriched versions. Their transformations are documented in the data and rights inventory and in the metadata of hsa-letters.json. Coordinate data come from Wikidata under CC0 and GeoNames under CC BY 4.0. The data and rights inventory lists attribution, modifications and licences for the bundled material.
Use CITATION.cff to cite the software. Machine-readable metadata are also available in codemeta.json.
Report problems or suggest improvements through GitHub Issues. CorrespExplorer is developed at Digital Humanities Craft.