A comprehensive AI analysis pipeline that analyzes AI news, benchmarks, and trends. Full explainer video is available on YouTube
This agent demonstrates a complex 5-agent sequential pipeline that:
- Fetches the latest AI news from Twitter/X using Exa search
- Retrieves AI benchmarks and analysis using Tavily search
- Scrapes and processes data from Nebius Token Factory Studio using Firecrawl
- Synthesizes and structures this information into a comprehensive analysis
- Analyzes AI trends and provides specific Nebius model recommendations
Uses a 5-agent sequential pipeline:
- ExaAgent: Fetches latest AI news from Twitter/X
- TavilyAgent: Retrieves AI benchmarks and analysis
- SummaryAgent: Combines and formats information from the first two agents
- FirecrawlAgent: Scrapes Nebius Studio website for model information
- AnalysisAgent: Performs deep analysis using Llama-3.1-Nemotron-Ultra-253B model
- Set up environment variables:
cp .env.example .env- Edit the
.envfile with your API keys:
NEBIUS_API_KEY="your_nebius_api_key_here"
NEBIUS_API_BASE="https://api.studio.nebius.ai/v1"
EXA_API_KEY="your_exa_api_key_here"
TAVILY_API_KEY="your_tavily_api_key_here"
FIRECRAWL_API_KEY="your_firecrawl_api_key_here"
Run with ADK CLI:
# Terminal - Run directly in the terminal
adk run analyzer_agent
# Dev UI - Visual interface for testing and debugging
adk web- Nebius Token Factory - For LLM inference
- Exa - For AI news search
- Tavily - For specialized search
- Firecrawl - For web scraping