Development · Web app
Multi-agent YouTube Video Analyst
End-to-end tutorial building a YouTube analyst using CrewAI, DeepSeek-R1, and Bright Data
A hands-on guide and reference architecture for building an AI-powered YouTube video trend analyst. It combines Bright Data for video scraping, a local DeepSeek-R1 via Ollama for LLM inference, and CrewAI to coordinate transcript analysis and response generation. A Streamlit interface provides the frontend for interacting with the multi-agent system.
What it's for
- Scrape and analyze video transcripts from targeted YouTube channels over specific date ranges
- Identify emerging topics and trends across competitor YouTube channels
- Learn how to connect local LLMs via Ollama to CrewAI agent workflows
Features
- Web scraping integration via Bright Data API specifying channels, dates, and country
- Local model inference running DeepSeek-R1 through Ollama
- Two-agent CrewAI orchestration with a Transcript Analyst and a Response Generator
- Streamlit front-end UI for triggering runs and reviewing output
- Complete code repository provided via GitHub
Pricing
Free tutorial and open-source code; requires Bright Data API credits and local hardware for Ollama
Details
- Type
- Web app
- Category
- Development
- Topic
- Agent Tutorials
- Open source
- Yes
- Platforms
- Web, Command line
- For
- AI engineers, data scientists, and developers building autonomous research agents
- Tags
- crewai, deepseek r1, ollama, bright data, streamlit, youtube scraping
Why it was saved
- @akshay_pachaar: The code steps (scrape → LLM setup → analyst agent → response agent → crew) are the tutorial the demo advertises; keeping the thread's build steps with the demo.
Related
An opinionated reference architecture for building full-stack open-source AI applications
MCP server providing structural AST indexing and semantic navigation for coding agents
Proxy Claude Code and coding agents through free, local, or alternative LLMs.
Open-source implementation recreating ChatGPT's interactive rich UI components
Updated