# How to Vibe Code Your Own ResearchRabbit (and Stop Paying for It) > Follow your curiosity with visual literature discovery - Site: https://researchrabbit.ai - Category: Academic Research & Literature Discovery - Verdict: **Serious undertaking** (42/100 vibecodeable) - Estimated effort: 4-6 weeks of part-time development ## Verdict Build a local personal citation explorer using OpenAlex and Cytoscape, but expect weeks of work on graph rendering and data pipelines. Replicating ResearchRabbit for personal use means you aren't cloning their private 310M-article database from scratch; you would query public scholarly APIs like OpenAlex or Semantic Scholar. However, the engineering challenge shifts to client-side performance: rendering an interactive citation network with thousands of nodes, smooth physics layouts, and responsive topic clustering in a web browser requires serious frontend architecture. Furthermore, syncing collections cleanly with Zotero involves navigating rigid bibliographic schemas and rate limits. ### What you can't replicate - ResearchRabbit's active user community and shared collection network - Proprietary recommendation model tuning trained on millions of real researcher exploration paths ## What it does An interactive bibliographic discovery engine that maps citation networks, co-authorship graphs, and research genealogies from seed papers. ### Core features - Seed paper input and bibliographic metadata ingestion - Interactive citation network graph visualization - Co-authorship and research genealogy mapping - Topic clustering and paper recommendation algorithms - Project collection management and folder organization - Zotero reference manager bi-directional sync ## The business ### Pricing - Free: $0 — Essential literature discovery - ResearchRabbit+: $10/mo — Large-scale academic reviews - Institution: Custom — Enterprise-scale university access Founded 2020. Team size: 10-15. ## The hard parts - Indexing and querying 310+ million academic articles with cross-referenced citation links - Rendering smooth client-side network graphs with thousands of interactive nodes and edges - Designing recommendation algorithms that accurately map topic proximity from seed collections - Maintaining stable two-way synchronization APIs with third-party reference managers like Zotero ## How to vibe code ResearchRabbit ### Prerequisites - Node.js (free): Runtime environment for building and executing the web application stack. - GitHub (free): Source code repository and deployment integration. - OpenAlex API / Semantic Scholar API (free): Free scholarly metadata source replacing the proprietary 310M article index. ### Recommended AI tools - Claude Code: Autonomous agentic coding tool for scaffolding complex graph visualization components and data ingestion loops. - Cursor: AI code editor for refining interactive graph UI elements and component states. ### Stack - Frontend: Next.js - Backend: Next.js API routes - Database: Turso - Auth: better-auth - Payments: none - Other: Cytoscape.js or Sigma.js for graph visualization, Vercel AI SDK ### Hosting - Vercel (Hosting the Next.js frontend and serverless API endpoints.): $0/mo - Cloudflare (Serverless edge database hosting via Turso/D1 integration.): $0/mo ### Build guide 1. **Project Scaffolding & Database Setup** — Initialize the Next.js project with TypeScript, configure Tailwind CSS, set up Turso SQLite for local collections, and integrate better-auth for single-user authentication. ``` Create a new Next.js 16 project with TypeScript and Tailwind CSS. Configure better-auth with SQLite via Turso to handle user sessions and project collections. Create database schemas for projects, saved papers, and seed collections with proper foreign keys and indexes. Write a test script to verify database connectivity and basic CRUD operations. ``` 2. **Scholarly API Integration Client** — Build a robust TypeScript service layer to query the OpenAlex and Semantic Scholar APIs for paper metadata, references, citations, and author profiles. ``` Create a scholarly data service in TypeScript that interfaces with the OpenAlex API and Semantic Scholar API. Implement methods to search papers by keywords, fetch paper metadata by DOI/ID, and retrieve inbound/outbound citation graphs for a given list of seed paper IDs. Implement caching mechanisms to prevent rate limiting and handle API error responses gracefully. ``` 3. **Interactive Citation Graph Visualization** — Implement an interactive node-link network visualization component using Cytoscape.js or Sigma.js to render papers, citations, and co-authorship links smoothly. ``` Build an interactive graph visualization component using Cytoscape.js inside a Next.js client component. The graph must render papers as nodes and citations as directed edges, supporting zoom, pan, node selection, and physics layout algorithms (cose or fcose). When a node is clicked, display a detailed sidebar with paper metadata, abstract, authors, and options to add it to the active seed collection. ``` 4. **Recommendation & Topic Clustering Engine** — Develop heuristic and vector-based recommendation algorithms that suggest related papers and cluster collection items by topic proximity. ``` Implement a recommendation and topic clustering engine in the backend. Given a set of seed paper IDs in a collection, aggregate their citation neighborhoods, compute overlap scores to rank related works, and cluster papers based on shared references and keyword tags. Expose an API endpoint that returns clustered recommendation groups for the frontend visualization. ``` 5. **Collections, Projects, and Zotero Export** — Build UI and backend logic for organizing papers into named projects and collections, with export functionality compatible with Zotero reference formats (BibTeX/RIS). ``` Build a project and collection management dashboard in Next.js. Users can create multiple research projects, add papers from the graph view into specific collections, add custom notes, and export selected collections in BibTeX and RIS formats for direct import into Zotero. Ensure responsive layout and clean state management. ``` ### Cost vs paying **Starting costs (one-time):** - AI Coding Assistant (Cursor/Claude Pro): $20 - Total: ~$20 one-time **Ongoing costs (monthly):** - Vercel Hobby & Turso Free Tier: $0/mo - OpenAlex / Semantic Scholar APIs: $0/mo - Total: $0/mo - Paying for the SaaS instead: $10/mo - Build time: 35-50 hours - AI tool credits: $20 - Break-even: Never (built for learning & personal research utility) ## Sources - [ResearchRabbit Official Website & Pricing](https://researchrabbit.ai) - [OpenAlex API Documentation](https://docs.openalex.org)