Project case study
Abstruck
From hours of watching to minutes of focused learning.
- Role
- Full-Stack Developer & Product Designer
- Status / year
- Active · 2026
Overview
Valuable knowledge increasingly lives inside lectures, meetings, podcasts, and tutorials. Returning to one explanation or decision often means replaying a long recording and manually finding the right moment.
Abstruck preserves spoken knowledge as a workspace people can search, study, and verify instead of producing another disposable AI summary.
The retrieval problem
Students replay lectures for one concept, professionals scrub through meetings for a decision, and self-learners lose useful ideas across long-form content. The information exists, but its original format makes retrieval slow.
A source-linked workspace
Users provide a lecture, meeting, podcast, or video. Abstruck turns it into a searchable transcript, an AI-generated summary with timestamp citations, organized notes, flashcards, and an interactive knowledge workspace.
Every generated insight links back to its original moment. Source traceability makes the output easier to trust, verify, and revisit.
Designing beyond the chatbot
The interface uses a calm editorial hierarchy so the product feels like a personal knowledge library rather than a chat window. Generated material is organized around reading, navigation, and recall instead of a stream of temporary responses.
- Searchable long-form transcripts
- Timestamp-cited summaries
- Structured notes and flashcards
- A persistent workspace for returning to source material
What the project demonstrates
Abstruck demonstrates product reasoning for applied AI: choosing a real retrieval problem, keeping generated claims connected to evidence, and shaping several processing outputs into one understandable learning workflow.
Technology
- Next.js
- TypeScript
- Auth0
- YouTube API
- AssemblyAI