Run an analysis
Ready.
Result
Placement
—Ideology map
—Same two axes, finer regions. The badge names the cell the account lands in.
Reading
—
Themes
Evidence
Caveats
—
How this is computed
The app pulls the most recent public posts from the account through the FxTwitter read API, trims them to a token budget, and sends them to a Groq-hosted OpenAI open-weight model with the prompt “read the tweets from this user, provide the political alignment of this user”. The model returns two axes in the range −1 to +1: economic (left to right) and social (libertarian to authoritarian). The marker is placed from those two numbers.
This is a heuristic reading of recent public text, not a measurement of a person's beliefs. Retweets and quotes are included and counted as signal. Short samples, sarcasm, quote-tweets and non-political accounts all produce weak placements, which is what the confidence score is for.