AI & Model Disclosure
Last updated: April 19, 2026
Presidentializer uses machine learning models to generate its output. This page explains how those models work and what data they were trained on.
Capitalization Model
The core feature — Trump-style capitalization — is powered by a Conditional Random Field (CRF) model trained on publicly available posts from Truth Social, sourced from the CNN Truth Social archive. The model predicts whether each word should be lowercase, Title Case, or ALL CAPS based on linguistic features including part of speech, named entity type, sentiment, and position in the post.
The model is retrained periodically as new posts become available. It is hosted on Hugging Face Spaces and runs inference on every converter request.
Nickname Generation (Optional Feature)
When the nickname toggle is enabled, the service uses the Anthropic Claude APIto generate nicknames for unrecognized people mentioned in your text, when sentiment toward them is negative. Your input text is sent to Anthropic's API for this purpose. Anthropic's Privacy Policy applies to that processing.
Known political figures use a curated static dictionary and do not involve the Claude API.
Accuracy & Limitations
- Model outputs are probabilistic and may not match the style in every case
- The model reflects patterns in training data and may reinforce biases present in that corpus
- AI-generated nicknames may be inaccurate, unexpected, or offensive
- Outputs should not be interpreted as factual statements
Training Data
The model was trained on publicly available social media posts. We do not claim copyright over third-party content used for training. The model itself is an original work trained by the Presidentializer project.