How to Build Agentic Web Research Assistants | Live Demo & Code
summary
Demo builds an Event Finder web research agent that searches live sites iteratively, aggregates results, and returns a validated JSON list with name, start/end dates, city, country and URL; it runs in Blazor .NET 9, uses browser geolocation and domain allowlists/blocklists, and calls the REKA research API in an OpenAI-compatible HTTP format. Iterative search depth defaults around 10 (demo requested 3) and reasoning steps are returned for debugging; code and a GitHub repo are published and the demo works live. structured JSON responses
AI-written summary · about
chapters (9)
- 0:00Agentic web research assistants
- 1:06Live demo: event finder agent
- 9:12Reka AI models and APIs
- 13:48Reka Research benchmarks
- 14:30Code structure breakdown
- 20:16API response and reasoning
- 23:51Geolocation and prompts
- 24:53Getting Reka AI API key
- 27:23Q&A
related talks
21mBring AI-Based Search to Your Web App
23mWebAI: Building the next user experience in the browser
59mHow to add genuinely useful AI to your webapp (not just chatbots)
01h 30mHow to build AI Agents: A step by step guide to creating AI agents
44mSupercharge your data with Azure AI Search and OpenAI
40mAI Agents in Practice
20mAgentic Excellence: Mastering AI Agent Evals w/ Azure AI Evaluation SDK
40mBuilding Intelligent Search Systems With Agentic RAG and OpenSearch