UPDATE 26/05/2026
I’ve updated this to 4 stars, as I spent the weekend working on it. With the help of Claude and the documentation, I found a workflow that has now resolved one of the scraping issues (Amazon).

The Notte workflow:
Use the CLI to explore interactively ← learn how the site works
Set up authentication if needed ← create a profile
Test interactions ← see what selectors/steps are needed
Export workflow code ← get a starting point
Adapt for functions ← use what we learned and the Function API

I’ve left off one star as I’m still testing. Also, when you create a function via the internal AI process, you can’t see the code in the deployed function. If it doesn’t work, you’re stuck and can’t edit it. If you manually create the functions, you can update them as needed.

PREVIOUS REVIEW

I've tried it on a couple of examples, and I've also implemented the API and CLI in a cloud computer called ZoComputer. And I've also tried using the interface directly in Notte. Unfortunately, in my test, it wasn't able to extract full information from eBay. However, I was able to get it to do it sort of one at a time through the CLI, but when going through the interface, it just didn't seem to want to do it properly. I was able to use BrowserAct to scrape eBay listings, so I'm not sure.

I also wanted to scrape some reviews from iHerb, and admittedly it does have anti-bot measures, but yeah, it failed at that as well. However, I've used another tool called Chat4Data, which is an extension for the web browser, and it's able to scrape all the reviews, no problem. Admittedly, BrowserAct does have this problem as well, so it's not just one tool. But I can only evaluate tools that I've got against each other and rate them as appropriate. Also, looking at the billing, it's burnt $8 of LLM and I didn't get any data, so, alright, for testing maybe, but at the end of the day, that's a lot to burn and not get any proper result.

I've also used Perplexity's Comet. While it's not scraping data and saving it, it can actually go through, say, all the iHerb reviews, and it can actually pick out all the positive and negative reviews, and I can ask it to isolate certain reviews and it will go and search. So, even though I'm not directly scraping data to download it, it is able to go and view that data and extract the data in terms of a summary, etc.

So I will keep testing this tool, but for now I have to give it a middle‑of‑the‑road rating because I'm not quite sure—at least with the AppSumo deal—what benefit it has, since it seems unable to scrape certain sites.