All guides
Research5 min read
Research a topic across YouTube — without 12 tabs
Compare tutorials, roundups, and opinion pieces on one subject. Rank, summarize, find consensus.
What's the problem?
Most people still leave chat to watch, scrub, and paste notes by hand. Here's the shift.
| Before | After Tubask |
|---|---|
| Twelve tabs, three half-watched videos, and a notes doc that still doesn't answer your question. | One research prompt. Tubask searches, ranks, and synthesizes — you stay in the conversation. |
How to do it
Three steps. Works in Claude, Cursor, or any MCP client connected to Tubask.
- 1
Frame the research question
Be specific: "best MCP OAuth tutorials" beats "MCP stuff." Add recency or view count if it matters.
- 2
Ask for rank + synthesis
Request top N results with channel names, then a paragraph on consensus and disagreements.
- 3
Drill into winners
Pick 1–2 from the list and ask for summarize_video on each — still in the same thread.
Ready to try this?
Connect Tubask in ~2 min, then paste your first YouTube link.
Copy-paste prompts
Replace VIDEO_ID or @handle with yours. Click Copy on any prompt.
Topic discovery
Prompt
Find the best recent YouTube tutorials on MCP OAuth. Rank by view count and list the top 8 with channel names and one-line summaries.
Compare approaches
Prompt
Research videos comparing "vector databases" vs "keyword search" for RAG. Summarize the consensus from the top results and note where experts disagree.
Recency-focused scan
Prompt
What are people publishing about AI agents on YouTube in the last 30 days? Top 10 by relevance, skip shorts.
Tips
- youtube_query with goal=research handles search and ranking.
- Ask for view counts or publish dates when comparing tutorial quality.
- Follow next_steps[] in the response — Tubask tells your assistant what to call next.
Common questions
Summary
Key takeaways
- Frame a specific research question — "best X tutorials" beats "X stuff."
- Ask for rank + synthesis in one prompt, then drill into 1–2 winners.
- Follow next_steps[] in the response for the recommended follow-up tool call.
Continue reading
More research outcomes.