ScholarAI
Deep Research Assistant
Overview
ScholarAI transforms the research process by combining agentic AI workflows with robust fault tolerance. The system leverages Temporal for orchestrating complex, long-running research tasks while Pydantic ensures data integrity throughout the pipeline. By integrating multiple AI providers and the Exa API for live web research, ScholarAI enables researchers to synthesize information from both private documents and public sources seamlessly.
Key Highlights
Engineered a fault-tolerant agentic workflow using Temporal via Pydantic AI to manage long-running research tasks, utilizing Pydantic for strict schema validation and structured output generation.
Enabled private deep research on user-uploaded files and live web data (via Exa API) by integrating OpenAI & Moonshot Kimi to synthesize contextual summaries.
Built a rich-text editor capability allowing the AI to generate and insert interactive elements like charts, tables, and diagrams directly into the user's drafting workspace.
Live demo coming soon