Read it laterCAN I VIBECODE
Genei?
BUILD FIRST
- One specific customer type and one primary use case
- A vertical slice that completes the main job without manual workarounds
- Visible limits, review steps, and a fallback when an external dependency fails
LEAVE FOR LATER
- Broad third-party integrations
- High-scale collaboration, syncing, and permissions
- Automation that can cause harm without human review
RESEARCH NOTES
What this tool actually does.
Genei (genei.io) is an AI-powered research and note-taking web app positioned as a 'read it later' and summarisation tool. Users import PDFs and webpages (via upload or a Chrome extension that saves pages for later reading), organise documents into projects/folders, and receive AI-generated summaries as hyperlinked bullet points that jump to the source passage when clicked. Core features: keyword extraction with definitions (named entity recognition), semantic/query-based search and question answering within and across documents, multi-document summarisation (Pro plan), AI-assisted note-taking with one-click add-to-notes from summaries/keywords, highlighting and comments, custom tags/definitions, citation management and reference generation, paraphrasing/rephrasing (Pro), and export. Summaries can be edited and the algorithm reportedly learns from those edits. Marketing claims include fa
RESEARCHED BUILD PROMPT
Give this to your coding agent.
Build a focused replacement for Genei, an AI-powered 'read it later' research and note-taking web app. Workflow: a user saves webpages (via a bookmarklet or browser extension) or uploads PDFs; documents land in a reading list organised into projects/folders; the app ingests each document, extracts keywords with definitions, and generates a summary as hyperlinked bullet points; clicking a bullet scrolls to the source passage; the user highlights passages, adds them to notes with one click, tags terms with custom definitions, and can ask questions or search semantically across one document or a whole project; notes can be exported and citations/references generated from linked passages.
User-facing requirements: PDF upload and URL import; reading list with projects/folders; per-document AI summary (bullet points linked to source positions); keyword extraction with definitions; semantic search and question answering scoped to a document or project; highlight/annotate; one-click add-to-notes; notes editor with Markdown export; citation/reference generation; a bookmarklet or extension to save pages for later.
Technical requirements: backend that fetches and parses webpages (readability extraction) and PDFs (pdf.js or similar); chunking and embedding pipeline (OpenAI embeddings or a local model) stored in a vector DB (pgvector or Qdrant) for semantic search and QA; an LLM summarisation service with prompt-based bullet generation; a split-pane reader/notes frontend (React/Next.js); async job queue for ingestion; auth and per-user storage.
Implementation approach: Next.js + Postgres/pgvector + a queue (e.g., BullMQ) + an LLM API; expose summarisation and QA as separate endpoints so models can be swapped.
Do not copy Genei's brand, logo, proprietary assets, or UI; build an original interface. Deliberately out of scope: mobile apps, team collaboration, a full browser extension (bookmarklet only), offline mode, and multi-document cross-summarisation (note as future work).
Include a README covering setup (env vars for DB and LLM API keys, migration/seed steps, dev server), architecture overview, and important limitations (LLM cost/latency, hallucination risk in summaries and QA, PDF parsing edge cases, rate limits).