All projects

Case study · 03 / 04

Thesa

A thesis-writing workspace that keeps chapters and research sources together, with a cited AI source-chat preview.

Screenshot of Thesa, the AI thesis assistantIn progress

01

The problem

Keep the research question, chapter drafts, and source material together while making answers traceable to the documents being read.

02

The approach

  1. Built a chapter-based writing workspace with a rich-text editor, a five-chapter template, and citation-style preferences.
  2. Added document upload and reading for PDF, DOCX, and TXT files, including page navigation and text search.
  3. Connected a local-preview Gemini assistant to semantic and keyword search over selected sources; answers include a quotation and a file/page location.
  4. Added autosave, revision-conflict handling, browser-draft recovery, and JSON backups.

03

What it does

  • Start from a five-chapter thesis template or a blank outline, with APA, MLA, or IEEE preference
  • Rich-text chapter editor with footnotes, tables, find/replace, and word count
  • Upload and read PDF, DOCX, and TXT sources, with page navigation and text search
  • Preview an AI source assistant that answers from retrieved passages and links back to the source page
  • Autosave, revision-conflict recovery, browser drafts, and JSON chapter backups

04

The outcome

The local preview supports writing, source reading, and cited source Q&A. The assistant is a local preview; cloud source imports still need the latest database migration, and citation insertion plus DOCX/PDF export remain unfinished.

Next projectNetPathLocal learning app · In progress