Turn monitored YouTube channels into a checked idea briefing with rtrvr

Define the channels and the collection

Luke’s January demonstration uses rtrvr to monitor YouTube channels, extract video data into Google Sheets and support video-idea research. It connects browser use, cloud workflows and APIs rather than relying only on manually reading channel pages. The resulting ideas remain editorial suggestions, not a prediction that a new upload will perform similarly.

1. Open https://rtrvr.ai/?utm_source=ai-luke&utm_medium=yt_influencer&utm_campaign=2026-q1-product-review and identify the channels, period and fields the research needs. Use the recorded AILUKE code only after checking its current terms.
2. Follow the Sheets extraction section at 3:01 with a small sample. Compare the returned titles, links and supplied metrics with the actual channel pages before accepting the collection.

Check the sheet and the cloud run

The video moves to cloud research at 5:44 and daily workflows at 7:22. A saved schedule does not confirm that every intended channel was checked. Preserve enough collection context to tell a quiet day from an access failure or a duplicate row.

1. Open the sheet and inspect representative records, missing fields and repeated videos. Keep retrieval dates and source links beside the information the agent will later analyse.
2. Test one cloud run and compare it with the expected channel list. Record failures explicitly and review the timing, timezone and destination before enabling a recurring version.

Connect the idea agent to the saved evidence

The API-and-agent section starts at 9:46 and topic suggestions at 12:18. The description names n8n and Claude 4.5 and retains a workflow JSON resource at https://sharetext.io/00db5ba1. Its free-access claim involving a Gemini API key describes the recorded offer, not a guarantee of unlimited usage or zero provider cost.

1. Use the retained JSON with the matching setup shown in the video and inspect the sources and credentials it requires. Run a small agent question against the collected records before extending the workflow.
2. Ask for the source videos behind each suggested angle and compare them with the sheet. Keep observed performance separate from your own choice of topic, title and audience promise, then save the reviewed ideas with their evidence.

The takeaway

Make every proposed idea traceable to the monitored data. Validate extraction, scheduling and the agent’s interpretation separately before treating the research as a dependable routine.