[Journey] Building a Custom Antidetect Browser & AI Automation

lilithark20911 said:

Hi BHW.

I'm a software dev, and I'm spending my nights and weekends after work building a custom antidetect browser from scratch.

I've been working on an AI automation setup to manage a few thousand social media accounts. But the commercial antidetect tools are just way too expensive. Paying a monthly fee per profile eats your margins instantly.

Even the providers that claim to offer "unlimited" profiles usually throttle you with hidden rules, like capping how many profiles you can create or open in a single day. It’s super frustrating when you're trying to scale. Plus, I really don't like the security risk of uploading all my cookies, proxies, and session data to their cloud servers. If they get breached, we lose everything.

So I decided to just build my own by modifying Chromium at the engine level.

The goal is to make a browser with absolutely zero limits—create or open as many profiles as your machine can handle. For the storage, I'm setting it up so you bring your own. You can save the profile data locally or just throw the database on a cheap, low-end VPS so you fully control it.

I'm building this to run my own bot fleet, but I plan to share the tool with everyone here once it's solid. We shouldn't have to pay crazy monthly fees or risk our data just to run our accounts safely.

I'll use this thread to post my progress, the technical headaches, and how the ban-testing goes.

Cheers.

Click to expand...

samratfkt said:

Using AI? if yes then which one you are preffering these days?

Click to expand...

kalpana007 said:

How much will be the cost to run 100 reddit accounts with this tool?

Click to expand...

Elon Snup said:

Building from scratch makes total sense at that scale. The per-profile cost on commercial tools literally caps your margins before you even start, and the unlimited tiers always have hidden throttles buried in the ToS that you only discover when it matters most.

The hardest part is not the Chromium fork itself - it is fingerprint persistence across sessions. Canvas hash, WebGL renderer string, AudioContext fingerprint, and the font enumeration output all need to stay identical every time the same profile reopens. If anything drifts between sessions, the platform detects a new device and starts the trust clock over from zero. Most open-source antidetect projects get this wrong and just regenerate the fingerprint on each launch.

WebRTC is the other one that bites people even when everything else looks clean. The local IP can leak through media API calls regardless of what proxy is configured at the browser level. You need to patch the WebRTC stack or disable it on a per-profile basis.

For the AI side - are you using the LLM to make decisions about what actions to take, or more for content generation and response handling? And are you patching Chromium at source level or doing runtime injection?

Click to expand...