May 8
Loading video player...
CEREBRO is a multi-agent AI Research Assistant built with CrewAI and n8n for INFO 7375 - Building Agentic Systems at Northeastern University. Given any research topic, CEREBRO automatically plans research questions, searches the web and 16 academic databases simultaneously, and synthesizes a structured, citation-rich report in 3-5 minutes. System Highlights: - 3 Agents: Research Coordinator (controller), Search Specialist, Synthesis Analyst - 4 Tools: SerperDev Web Search, Website Scraper, File I/O, and CEREBRO custom tool - CEREBRO Custom Tool: 940-line, 16-database academic meta-search engine (ArXiv, Semantic Scholar, CrossRef, OpenAlex, PubMed, Zenodo, CORE, DBLP, and 8 more) - Hybrid Architecture: CrewAI for agent orchestration + n8n for workflow automation - 3 Execution Modes: Streamlit Web UI, CLI, Webhook API - 27 Tests: 13 unit + 9 integration + 5 end-to-end (all passing) - Baseline validated: +27% source relevance over web-only search Built by: Aravind Balaji Course: INFO 7375 - Building Agentic Systems Professor: Nik Bear Brown University: Northeastern University Date: April 2026 GitHub: @B98/Cerebro Play in 2x - 2.5x speed for less than 5 minute video. #AI #AgenticAI #CrewAI #n8n #MultiAgent #ResearchAssistant #NortheasternUniversity #LLM #Python