Introducing Demonstrate Mode: For Absolute Precision
Record your manual browser actions and Notte generates production-ready deterministic code. For when you need pixel-perfect execution, not AI guesswork.
LLM agents are incredible at figuring out "how" to do a task. But sometimes, you don't need discovery—you need specific, pixel-perfect execution.
If you are building a complex checkout flow or navigating a legacy internal tool, you can't afford for the agent to "guess" the path. You need it to click exactly where you would click.
Enter Demonstrate Mode.
We built a high-fidelity session recorder that turns your manual actions into production-ready code.
How it works:
- You Drive: Launch a live browser session in the Notte IDE.
- You Execute: Click through the flow, handle the edge cases, and select exactly the data you need.
- We Code: Notte records your intent and generates the deterministic script (selectors, waits, interactions) in real-time.
The New Hierarchy of Automation
We are building the complete IDE for browser automation. You now have two distinct ways to build:
- Agent Mode: High leverage. Great for exploration, bootstrapping, and fuzzy tasks.
- Demonstrate Mode: High precision. Great for complex flows where you know exactly what needs to happen.
Don't spend 20 minutes prompting an agent to find a button you can see in 2 seconds. Just click it, capture the code, and ship.
Get Started
Click through browser flows today: console.notte.cc
Frequently asked questions
What is Demonstrate Mode in Notte?
A high-fidelity session recorder that turns your manual browser actions into production-ready deterministic code. You launch a live browser session in the Notte IDE, click through the flow yourself, and Notte generates the script with selectors, waits, and interactions in real time.
When should I record a browser flow instead of prompting an AI agent?
When you need pixel-perfect execution rather than discovery, such as a complex checkout flow or a legacy internal tool where the agent cannot afford to guess the path. If you can see the button in two seconds, clicking it and capturing the code beats twenty minutes of prompting.
What is the difference between Agent Mode and Demonstrate Mode?
Agent Mode is high leverage: good for exploration, bootstrapping, and fuzzy tasks. Demonstrate Mode is high precision: good for complex flows where you know exactly what needs to happen.
Does Demonstrate Mode generate code I can run in production?
Yes. It records your intent and generates a deterministic script with selectors, waits, and interactions that is meant to ship as production code, not a rough draft for an agent to reinterpret.