thanks! i asked my teammate working on bot detection to respond, he can better answer this!

Hi Bronco, Omar here, a Lead Platform Engineer at Intuned.

> I've tried adding browser automation outside of the single click from the extension, and I'm immediately met with bot detection.

Can you explain how you write your automation? How do you do the click from the extension? Do you use CDP input commands to perform the click (`Input.dispatchMouseEvent`), or do you execute JS code and click the button using `element.click`? using CDP will give you much better score than Executing JS to click a captcha button. JS Execution can be easily detected by any bot detection provider. On the other hand, using CDP to click can mimic actual mouse movement and will make it much harder to detect abnormalities. and the click event will have `isTrusted=true` and handle detection methods related to monitoring JS executions on page.

> Do modified browsers drive the overall bot detection heuristic low enough that the cloud IP address itself isn't a red flag? I've seen Camoufox and will try it at some point. What other options are available to drive down the overall "score" so I can still automate the browser but keep the latency low?

Modified browsers reduce your bot score a lot, and Camoufox is a great option to test out. Will it work? It depends on how the website has set up their bot detection. using modified browser is a must for use case.

At Intuned, we use our own internal forked Chrome to hide the most popular signals, and a lot of the time, the browser alone without a residential proxy can help us bypass most websites, but not all of them(IP reputation has very high heuristic value).

I can't give exact recommendation on what will work with you for sure since each website has it's own ways to handle bot detection , one recommendation I can give is to try to use packages like patchright they can help alot and hide alot of popular signals.

Another recommendation I can make is to try using intuned agent and ask it to help you find a way to bypass bot detection on that website. If you can handle it using network interception or some other scraping technique, the agent is really good in these cases and knows most of the used scraping techniques.

thanks! give us a try next time and checkout fix with ai when the scraper breaks!

Great question - we tried so many things and overtime, we ended up doing 2 things - custom chromium build and extension to solve captchas. In our docs, checkout stealth mode - https://intunedhq.com/docs/main/02-features/stealth-mode-cap...

Also, one of our engineers did a write up on bot detection systems and how they work - https://intunedhq.com/blog/how-bot-detection-works

If you think about "price", "speed" and "accuracy (reliability/quality)", our bet is that models won't hit those 3 together. So you won't get a model that is very fast, very cheap and very accurate anytime soon.

Also, imagine that you have a case where you want to scrape 10,000 records from a website, why have AI navigate to every page to do this? why not write the code, run it, and get consistent and fast result? its also predictable, if it messes up, you know what happened and you can trace it to the exact line of code.

this is not meant as personal assistant to do a task like this - the idea is that if you want to build a travel assistant and you want to do an integration with a travel booking service using browser automation - you can use Intuned to build this integration and expose an API to do search and one to do a booking, etc. Does this make sense?

We actually went through 1 hard pivot only, the reset is more about framing the problem as we dug deeper and understood the customers and issue more. As an example, "Intuned Auth Sessions" is just a feature that we still support today! But you are right, being a founder is not easy and the hardest part is figuring out what to build, does it make sense to keep going or should you stop - those are questions I still struggle with until today.

For your question about how is this different - I think if you dig into those product you will see that our focus is different, many of the companies mentioned are focused on powering agents via APIs, some are focused on enabling users to use AI at runtime, we do feel that our product is somewhat differentiated - the closest one is possibly Reworkd and I would still say the product is somewhat different. Now, the hardest part is actually commenting this with customers and the market in general - and there, we have a lot to figure out!

For robots.txt and user-agents question, we think of ourselves as providing infrastructure and flexibility for our customers to do what they want - we do encourage in our docs that they respect robots.txt but we don't enforce it on a platform level.

Appreciate you taking the time to leave this comment - very thoughtful

thanks for the kind words! We have been seeing this pattern of customers wanting full service since we launch. Let's see how it goes!

thank you!

thank you!

We actually do support CDP directly as well. As for why playwright, playwright provides a lot of benefits that we like in terms of being code - like auto-waiting, good syntax for a lot of operations, agents can use playwright pretty well and playwright is generally like by the community we target.

If a customer, doesn't want to use playwright, they don't have to given CDP but we most of our templates use playwright.

Yes, we we did our YC and social launches, we had few companies sign up and we have been building with them sense. For some of them, we have enabled them to run 1000+ scrapers which would have been very hard without Intuned. Some of these have been with us for 2+ years! What our users love is the agent and the idea that we are not using AI at runtime which improves reliability.

definitely came up multiple times. Next week, we are releasing our codex/claude plugin, so you will be able to use codex to create the projects and deploy them to Intuned.

Intuned as a platform to deploy browser automation adds a lot - anti-bot detection, jobs, observability and more.

So far its cost me $2.27 to submit a contact form 3 times - why is this better than a captcha solver with human solves at 1000 per $2?

On your automation, your tool fed back to me as follows after 3 submissions:

> The CAPTCHA is persistently blocking now — Prosopo's widget appears to have flagged the session/IP due to the repeated submissions. The checkbox won't reset this time. This is expected behavior from their bot protection product. To submit again, you'd likely need to wait a while for the rate limit to clear, or submit manually from your own browser.

This never made it into prod since the scale was small, but one of the favorite leaks I found when working on bot detection was browsers which generated the same random numbers. Presumably because they were being init to the same VM snapshot and therefore the same random number state.

What is the advantage of your product over having Codex generate a script using something like https://github.com/D4Vinci/Scrapling?

I see I'm already using a FOSS project for anti-bot

for jobs/durability/obs i have sqlite and had codex generate an ugly but functional dashboard

im just curious to know what intune does that is different

The cost is AI cost for using the agent - not captcha cost. Usually, you would write the project and then call it via API - instead of asking the agent to do the action more than 1 time. Considering using the web task API for this use case.

our product is infra + agent. You can use codex or other agent to generate the code. We actually have a cli that allows you to deploy projects to our infrastructure.

We are actually working on open sourcing a plugin that you can use with any coding harness!

oh neat!

other than the stuff I mentioned, it’s the deep integration between the agent and platform. Because we have obserabiltity, you can open a failed run, and with a click of a button fix it. you can also enable self healing on a project which puts it on autopilot. we have a lot to improve on but if you have 10+ active scrapes, intuned as a package saves users time and pain.

are there code or demo i can view

that sounds interesting

fix with ai - https://youtu.be/e4R4hLdHBro self-healing demo - https://youtu.be/IVHIXw0lYMs

I am happy to give you a demo over a call as well