📚 Wikipedia Research Agent
Point the agent at a seed article and a topic to research.
The actual app, as it opens in NodeTool — built with the visual App Builder, no code.
What you put in
- textSeed URL
- textTopic
What you get out
- TextFindings
How it works
- Step 1
Fill in a few fields
Plain-language inputs — seed url, topic. No settings, no jargon.
- Step 2
Click “Research”
The workflow runs behind the scenes — on your machine or with your own provider keys. You pay providers directly, or run local models for free.
- Step 3
Watch results stream in
Results appear live in the app: findings.
It's an app until you want more
Every mini app is a real NodeTool workflow underneath. Open the canvas to swap models, change prompts, or add steps — then publish your remix as a new app for your team.
Explore the Wikipedia Agent workflowMore mini apps
Paste a paper or abstract and extract the key findings.
One topic in — a keyword-targeted article batch out, with hero images.
An agent scours YouTube and distills what actually matters.
Turn one product photo and one offer into a batch of ready-to-test video ads.
Ask anything — an agent runs the Google searches and reports back.
Describe a scene out loud — get it back as a picture.