An early summarization prototype
Castela
A historical Flask-based experiment in text summarization.
I designed and built this project independently, using my own resources.
Explore the project
Public repository
Technical documentation
Archived Flask prototype. The application source and dependencies are preserved.
Technologies
01 / Context
The problem behind the project
An early prototype provided a small web surface for experimenting with summarization.
Castela is a small early experiment in turning a web page into a summary. It predates the later shared AI service and is intentionally shown as a prototype rather than being retroactively described as a complete AI platform.
02 / Implementation
Design and implementation
I explored summarization in a Flask application before the later shared AI-service work.
I built a Flask application that accepts a URL, downloads its HTML, extracts text with BeautifulSoup, runs the BART CNN summarization pipeline, and renders the summary back into the page. The retained implementation is a short synchronous request path, with model inference in the application process.
03 / Approach
Engineering decisions
Retain the experiment as history
The public source contains a Flask application and templates. Its historical scope is kept distinct from the current Mistify service.
04 / Evidence
Evidence and outcomes
The public application source directly shows the complete fetch → text extraction → summarization → template-render flow and its configured output-length bounds.
Source and technical notes
05 / Current scope
Where it stands
A historical Flask summarization prototype.
The prototype does not implement the durable queue, authenticated ingestion, source provenance, URL-fetch hardening, or long-document handling of a production editorial system. Its value here is the earlier experiment and the progression toward more explicit service boundaries.