Letters Promptotype

A Flask-based prototype for letter editions, employing DH Craft’s Promptotyping methodology.

This project serves as a promptotype—a prototype developed using AI-assisted prompt engineering—designed to facilitate the exploration and visualization of letter corpora in a structured XML TEI format.

Overview

The app provides: - a list of letters with status labels. - access to a single-letter view with multiple tabs: - Original text - Translation (if available) - TEI/XML Code view - Basic metadata - a timeline showing when letters were sent.

Data Sources

Currently, the prototype is based on letters from the Joseph von Hammer-Purgstall digital letter edition (coming soon on http://gams.uni-graz.at/hpe. However, the data can be replaced with letters from the Hugo Schuchardt Archive http://schuchardt.uni-graz.at (available in the schuchardt-data folder). Simply replace the dataset in the data/ folder and restart the app.

License:
The letter data and the app is licensed under Creative Commons BY-NC 4.0.

Technical Details

• Framework: Flask (Python)
• Hosting: Render.com
• Frontend: HTML, Bootstrap5
• Data Storage: JSON-based
• Dynamic Data Sources: Can be adapted to different letter collections with similar structure.
• Scalability: Successfully tested with ~500 letters.

How to Run Locally

Clone the Repository

sh git clone https://github.com/martinascholger/promptotype-letters.git cd promptotype-letters/letters

Install Dependencies

sh pip install -r requirements.txt

Run the Flask App

sh python app.py By default, the app runs at http://127.0.0.1:5000/.

Live Deployment

The app is deployed on Render.com, and the latest version can be accessed at: Live App URL

🔧 Development Process

The app was developed using Promptotyping, following these stages: 1. Prompt Engineering
- Designed structured prompts for AI-assisted interface generation using Claude 3.5 Sonnet & ChatGPT o1. See more on DH Craft’s Promptotyping Process Guidelines. 2. Setup & Environment
- Flask backend with JSON data storage. 3. Refinement
- Improved individual views (list view, letter detail, metadata, timeline).

Extending the Prototype

• Replace the data folder with other TEI-structured letters (e.g., from the Hugo Schuchardt Archive).
• Modify the Flask routes to adjust views.
• Implement additional filters for letter sorting and search.

Citation

To cite this project in your research, please refer to the CITATION.cff file. You can also use GitHub's citation feature by clicking on the "Cite this repository" button on our GitHub page.

References

Hurch, Bernhard (Hrsg.): Hugo Schuchardt Archiv. Graz: Institut für Sprachwissenschaft, Universität Graz. 2022. URL: http://schuchardt.uni-graz.at.

Höflechner, Walter; Scholger, Martina; Wagner Alexandra (Hrsg.) 2025: Joseph von Hammer-Purgstall. Korrespondenz. Graz. URL: https://gams.uni-graz.at/hpe (forthcoming).

Steiner, Christian / Pollin, Christopher (2024): Promptotyping: AI-Enhanced Rapid Prototyping for Data-Driven Applications. https://github.com/DigitalHumanitiesCraft/excellence/tree/main/promptotyping

Steiner, Christian, / Pollin, Christopher (2024): Promptotyping (v1.0.0). Zenodo. https://doi.org/10.5281/zenodo.14160876