Build a Production-Ready YouTube AI Agent with Temporal
We build a deep research agent over the DataTalks.Club podcast archive. We start by downloading and indexing YouTube transcripts. Then we turn that ingestion code into a durable Temporal workflow. Once the data is searchable, we build a Pydantic AI research agent. We add a summarization sub-agent for long transcripts and wrap the agent run in Temporal too.
Links
The main resources for this workshop:
The system you will build
The final system looks like this:
The ingestion side has the parts that usually fail in production. It makes network calls, runs proxies around YouTube's cloud-IP blocks, and writes to Elasticsearch. It also loops over many videos. Temporal gives that side retries, observability, and durable execution.
The agent side uses the indexed data to answer questions from the podcast archive. It uses Temporal again too, so long agent runs can survive failures.
Tutorial pages
- Overview and setup
- Part 1: Fetch one transcript
- Part 2: Run Elasticsearch
- Part 3: Discover podcast videos
- Part 4: Temporal motivation
- Part 5: Define the workflow
- Part 6: Create the agent project
- Part 7: Summarize long transcripts
- Part 8: Wrap the agent in Temporal
- Q&A: questions from the workshop
- What to improve next
- Appendix: file inventory