Scholar AI

ScholarAI is your AI-powered research companion—streamlining everything from literature search to contextual Q&A with memory, knowledge graphs, and intelligent summaries.

Comment

Inspiration

  • The academic research process is slow and fragmented—finding, reading, synthesizing, and revisiting literature all require manual effort.

  • Existing tools don’t retain your research “context,” making it hard to have meaningful, multi-step conversations with AI about your work.

  • U wanted a dashboard that not only summarizes papers, but also remembers your research journey, grows your knowledge graph, and supports natural follow-up Q&A.

What it does

  • Fetches and summarizes the top academic papers for your topic, all in one click.

  • Generates full research reports with word count and read time for easy sharing or review.

  • Enables contextual Q&A: You can chat with your AI research assistant, which remembers previous reports and questions—so follow-ups are grounded in your real work.

  • Stores your research history with persistent memory, enabling you to pick up where you left off.

  • Coming soon: Visual knowledge graph editing and peer collaboration.

How I built it

  • Frontend: Built with React and TailwindCSS, using a modern dark purple gradient theme for clarity and focus.

  • Backend: FastAPI server with CrewAI agents for literature search, report writing, and chat.

  • LLM & Tools: Uses GPT-4 for intelligent summarization, report generation, and chat, with SerperDev for scholarly search.

  • Memory: Integrates Mem0 AI for long-term and cross-session user context and chat history.

  • Persistence: All research topics, reports, and Q&A are saved per user to enable multi-turn, memory-rich sessions.

Challenges I ran into

  • Handling multi-turn, context-rich Q&A with persistent memory—getting the agent to truly “remember” past reports and use them in new conversations.

  • Adapting to new LLM APIs and memory storage requirements (e.g., Mem0 formats and agent IDs).

  • Fine-tuning UI/UX to keep the interface intuitive even as features grew (report, chat, knowledge graph previews, etc.).

  • Ensuring fast, reliable responses despite the complexity of chaining multiple AI agents and API calls.

Accomplishments that I am proud of

  • Built a real, context-aware chat system that actually grounds responses in your own research outputs.

  • Seamlessly linked literature search, markdown report generation, and contextual chat—all with persistent memory.

  • Designed an interface that is clean, modern, and “feels” like an all-in-one academic dashboard.

  • Created a solid base for knowledge graph and peer collaboration features.

What I learned

  • True research assistants need memory—one-off LLM chats are not enough for serious academic workflows.

  • Carefully designed memory and entity storage unlock richer, more human-like interactions with AI.

  • User experience matters: users want a simple, beautiful tool that “just works,” not a dozen disconnected scripts.

  • Collaboration and knowledge mapping are highly requested in academic communities.

What's next for ScholarAI

  • Knowledge Graph UI: Visualize, edit, and extend the web of your research topics, sources, and insights.

  • Editable Reports: Let users annotate, update, and expand reports directly in the dashboard.

  • Peer Collaboration: Enable real-time co-authoring and sharing of research with classmates or colleagues.

  • Reference Management: Organize, cite, and export your literature for academic writing.

  • Enhanced Memory: Further improve cross-session recall, summarization, and personalized suggestions.

Built With

Share this project:

Updates