Skip to content
All products
Applied NLPLive

Verity

Verity takes a piece of text and returns a credibility signal in real time. It was built to explore how far a disciplined, classical NLP pipeline — not a large model — could go on a well-defined classification task, deployed end to end as a real product rather than a notebook.

Core capabilities

  • TF-IDF vectorization tuned for short-form text
  • Logistic regression classifier trained on labeled news data
  • Flask API with request validation and structured responses
  • Static frontend calling the API directly, no server-rendered coupling

Frequently asked

Does Verity use a large language model?+

No. It deliberately uses a classical TF-IDF plus logistic regression pipeline, to see how far disciplined feature engineering goes before reaching for a larger model.

Is the source available?+

Yes, the backend is open source. The link is on this page.

Use cases

  • Quick credibility check before sharing an article
  • Reference implementation for lightweight text classification

Interested in something similar for your team?

Start a conversation