PandaStudio
Browse all docs

Find your next video idea from YouTube outliers in PandaStudio

PandaStudio's AI agent can research what to film next: it finds YouTube videos that got many times their own channel's usual views, works out why, and turns that into video ideas with titles, hooks and the outliers that prove each one.

An outlier is a video that beat its own channel's median views, so a 10x outlier got ten times what that channel normally gets. That means its topic, angle or title pulled people in beyond the channel's usual audience, which is the best evidence there is that an idea works. Ask the agent about your niche and it searches YouTube, reads every channel it finds, ranks their recent uploads by how far they beat their own normal, and spots the title patterns behind them. It can also look at a single channel (yours included), find the channels competing for the same viewers, read an outlier's transcript and top comments to see why it worked, and check how much people search for a phrase. Then it writes ideas, each with a title, the angle, a 15-second hook, thumbnail text and the outliers it is based on. You can keep a list of competitor channels and ask what is new with them any time. It uses public YouTube data, so you do not need to connect a YouTube account. Saved research is deleted after 30 days.

In the app

  1. 1Open the AI tab (or the floating chat bubble) in PandaStudio.
  2. 2Ask what to film next, naming your topic or a channel your viewers watch.
  3. 3Read the outliers and ideas it comes back with, and ask for more on the ones you like.
  4. 4Pick one and keep going in the same chat: script it, record it, then edit it.

Ask the AI

Open the AI chat in the editor and say any of these. The agent does the edits for you.

“What should I film next? My channel is about home espresso. Find outliers from the last few months.”
“Find the channels competing with @JeffSu and what's breaking out for them.”
“Why did this video do so well? https://youtube.com/watch?v=...”
“Track these five channels as competitors and tell me every week what broke out.”
“Will "I tried every budget espresso machine" work as a title? Check search demand.”

CLI & MCP

Every capability is a verb, reachable from the terminal or from an MCP agent (Claude Code, Cursor, Cline, Continue, Claude Desktop).

ideas.research
pandastudio ideas.research --query="home espresso, espresso machine review" --minMultiplier=3

Takes 20 to 40 seconds and runs as a job: wait for it, then read the results with ideas.get.

ideas.get
pandastudio ideas.get --researchId=<id> --onTopicOnly
ideas.channel
pandastudio ideas.channel --channel=JeffSu
ideas.similar
pandastudio ideas.similar --channel=JeffSu
ideas.video
pandastudio ideas.video --video=<url> --baseline --comments=100
ideas.transcript
pandastudio ideas.transcript --video=<url>
ideas.keywords
pandastudio ideas.keywords --seed="espresso machine"
ideas.demand
pandastudio ideas.demand --keywords='["budget espresso machine","espresso for beginners"]'
ideas.track
pandastudio ideas.track --channel=JeffSu

Then ideas.scan re-reads every tracked channel and lists what broke out. ideas.competitors shows the list.

Edit a video by asking AIRecord your screen and cameraAll features