# ai.txt — policy and discovery metadata for www.bastianportes.de
# This file is a machine-readable rights notice. It is not a copyright license.

version: 1.0
canonical: https://www.bastianportes.de/
language: de-DE
contact: mailto:bastian@servus-medientechnik.de
llms: https://www.bastianportes.de/llms.txt
sitemap: https://www.bastianportes.de/sitemap.xml
catalog: https://www.bastianportes.de/data/werke.json

[permissions]
search-indexing: allow
user-initiated-retrieval: allow
answer-engine-use: allow-with-attribution
quotation: statutory-limits-only
model-training: disallow
model-fine-tuning: disallow
dataset-creation: disallow
text-and-data-mining: disallow
synthetic-content-generation-from-works: disallow
full-text-reproduction: disallow
media-reproduction: disallow

[attribution]
required: true
preferred-name: Bastian Portes
preferred-source: https://www.bastianportes.de/

[training-restricted-paths]
path: /downloads/
path: /leseprobe-mordsmoor.html
path: /leseprobe-owm.html

[rights]
statement-de: Der Rechteinhaber behält sich Text und Data-Mining gemäß § 44b Abs. 3 UrhG ausdrücklich vor. Ohne vorherige schriftliche Zustimmung dürfen Texte, Leseproben, Bilder, Videos und sonstige Werke weder für Training oder Feinabstimmung generativer KI noch zur Erstellung von Datensätzen oder für vergleichbare maschinelle Auswertungen verwendet werden. Gesetzlich zwingende Ausnahmen bleiben unberührt.
statement-en: Rights for text and data mining are expressly reserved under Section 44b(3) of the German Copyright Act. No model training, fine-tuning, dataset creation or comparable machine analysis is permitted without prior written consent. Mandatory statutory exceptions remain unaffected.

[retrieval-authorized-summary-pages]
url: https://www.bastianportes.de/mordsmoor.html
url: https://www.bastianportes.de/offenes-weltmodell.html
url: https://www.bastianportes.de/kurzroman.html
url: https://www.bastianportes.de/autor.html
url: https://www.bastianportes.de/presse-film.html

[notes]
note: Retrieval permission applies to concrete user-initiated questions and requires clear source attribution.
note: Authorized reading samples remain copyrighted extracts and are not training data.
note: This policy supplements robots.txt and the copyright notice in the imprint.