Project Story: ScholarGen

About the Project

ScholarGen is an AI-powered academic research and writing application designed to help researchers move from a research idea to a complete, submission-ready academic paper, step by step.

The platform focuses on structured reasoning, academic rigor, and workflow guidance—rather than simple text generation. It is built to support complex academic outputs such as Master’s theses and Systematic Literature Reviews (SLRs).

The Problem Being Solved

Academic research writing is difficult not because of writing alone, but because of:

Poor structure and planning
Weak linkage between literature review, methodology, and results
Time-consuming formatting and citation management
Cognitive overload when managing large, multi-chapter documents

Most AI tools generate text in isolation, which often leads to:

Inconsistent chapters
Methodology that does not follow from the literature
Outputs that are difficult to defend academically

ScholarGen addresses this gap directly.

Core Insight

Good academic writing is not about better sentences—it’s about better structure and reasoning.

Academic research should be treated as a sequence of connected tasks, not a single text-generation prompt.

What ScholarGen Does

ScholarGen acts as a research workflow engine that:

Helps users define a clear research problem and objectives
Generates a structured research plan (chapters and sub-chapters)
Allows users to execute each section independently
Maintains logical continuity across the entire document
Produces properly formatted, citation-ready outputs

Each chapter builds on previous ones, ensuring that:

The literature review informs the methodology
The methodology aligns with the stated objectives
Results and discussion remain consistent with earlier reasoning

The Solution

ScholarGen provides an AI-guided, step-by-step academic workflow:

Project Configuration
Define field of study, academic level, language, and scope.
Research Definition
Clarify the problem statement, motivation, and objectives.
AI Research Planning
Generate a structured chapter and sub-chapter plan.
Controlled Execution
Write and refine each section individually, with AI assistance that understands prior context.
Finalization & Export
Generate submission-ready documents (DOCX or PDF) with consistent formatting and structure.

How We Built It

ScholarGen was built as a client-side Single Page Application (SPA), designed to be lightweight, responsive, and fully browser-based.

For academic research applications that require complex reasoning and high-quality generation, Gemini 3 Pro Preview was chosen as the AI model due to its strong long-context handling and academic coherence.

Technical Architecture Overview

Core Technology Stack

React (v19): UI component structure and application state
TypeScript: Type safety for complex research data structures
Tailwind CSS: Utility-first responsive styling (CDN-based)
FontAwesome: Interface iconography

AI Integration (The “Brain”)

The core intelligence is powered by the Google GenAI SDK (@google/genai).

Key AI services include:

Research Planning: Generates structured chapter/sub-chapter JSON trees
Content Generation: Writes academic text using prior chapters as context
Utilities: Summarization, paraphrasing, and humanizing tools
Error Handling: Graceful handling of rate limits and safety blocks

All AI logic is centralized in services/api.ts.

Application Architecture

The application follows a Linear Workflow combined with a Workspace model:

State Management:
The main App component (index.tsx) holds the current project as the single source of truth.
Phased Workflow:
- Configuration
- Definition
- Planning
- Execution
Key Components:
- ProjectDashboard: Manages saved projects in local storage
- ExecutionWorkspace: Main research environment
- SubChapterEditor: Writing surface with local undo/redo history

Data Persistence & Export

LocalStorage:
All project data is stored locally in the browser as serialized JSON, enabling a serverless architecture.
Document Export:
- docx: Generates Microsoft Word documents in-browser
- jspdf: Generates PDF files

Key Features

Auto-Save:
Uses a debounced save mechanism to prevent performance issues.
Versioning & Undo:
Each sub-chapter maintains its own history stack for undo and redo actions.
Serverless by Design:
The application runs entirely in the browser, simplifying deployment and scaling.

Use Cases

Master’s and postgraduate theses
Systematic Literature Reviews (SLRs)
Academic research proposals
Early-stage doctoral research
Research training and supervision support

Scalability

ScholarGen is designed to scale:

Technically: No backend dependency
Functionally: Easy to add new methodologies, citation styles, and AI models
Academically: Supports undergraduate to doctoral-level research

Challenges Faced

Key challenges included:

Maintaining logical continuity across AI-generated sections
Preventing context loss in long-form documents
Balancing automation with researcher control
Designing an interface that supports deep thinking, not distraction

What We Learned

Academic rigor comes from structure, not verbosity
Researchers need control, not just speed
AI is most effective when it supports reasoning rather than replaces it
Long-form academic work requires persistent context and planning

Vision

The long-term vision for ScholarGen is to become a trusted academic research companion that helps researchers think more clearly, structure stronger arguments, and produce defensible work faster.