What YouTube data teaches us
Outliers, niches, titles, thumbnails: analyses drawn from millions of videos, and how to use them.
4 min read
ChatGPT vs Claude for YouTube: which one is better for creators? (2026)
An honest comparison of ChatGPT and Claude for YouTube work: scripts, titles, research and data. Why neither knows what's performing this week on its own, and how connecting Claude to a YouTube database changes the answer.
Read the article5 min read
20 Claude prompts for YouTube research that use real data
Copy-paste Claude prompts for YouTube creators and agencies: video ideas from outliers, competitor analysis, title formats, niche research, comments and scripts. Each one works with real data once Claude is connected to Returbe's MCP.
4 min read
Videos about Claude are exploding on YouTube: the data and the angles that work
Videos about Claude and Claude Code are getting 20x to 870x their channel's usual views. We looked at the outliers of the last 90 days: which angles work, which title formats repeat, and how small channels are riding the topic.
5 min read
How to go viral on YouTube with Claude (a workflow built on real data)
Claude can't guess what will go viral, but it can read what already is. Here's a 5-step workflow that connects Claude to 1.5 million tracked YouTube videos to find outliers, steal the right format and pick a title, with real examples from November 2026.
5 min read
YouTube MCP server: connect Claude to real YouTube data (setup + 10 use cases)
What a YouTube MCP server is, why Claude needs one, how to connect Returbe's in 2 minutes in Claude.ai, Claude Desktop or Claude Code, and the 25 tools it gives Claude: outliers, competitors, niches, keyword demand, comments, transcripts.
4 min read
YouTube research for agencies: how to cut client research from days to an hour with Claude
How YouTube agencies and production studios use Claude connected to Returbe's MCP to find ideas, map competitors and build weekly client reports from real data, with the exact prompts and a workflow per client.