mistervanderbelt
Newbie
- Feb 21, 2025
- 8
- 8
Been down a rabbit hole over the past year trying to figure out why every automation setup built on the public Reddit API eventually underperforms, and wanted to share what I found because I don't see this discussed much here.
The public Reddit API is rate limited, heavily monitored, and most importantly it identifies itself. When you make requests through the official API, Reddit's servers know exactly what you are. You're not a user. You're a script. That context is baked into how your traffic is scored and how your accounts are treated.
Reddit's actual iOS app doesn't use the public API at all. It runs on a completely separate internal client layer with its own authentication system, its own token structure, its own request signatures. The iOS app has a two tier token system that the public API doesn't replicate. It sends device attestation headers. It passes a device fingerprint with every request. All of this is invisible if you've only ever worked with the public API.
What this means practically: any tool built on top of the public Reddit API is operating at a disadvantage by design. You're using the front door Reddit built for developers to read data, not the same channel their own app uses to behave as a real user.
When I realised this I stopped building on the API entirely and started looking at what the iOS app was actually sending at the protocol level. Captured the full client session from a real iOS device, every identifier, every header, every token, and built infrastructure to replicate that at scale.
The result is something that behaves, from Reddit's perspective, exactly like a genuine iPhone running the Reddit app. Not an API call. Not a browser. An iPhone.
For anyone doing serious volume on Reddit this distinction matters more than almost any other variable. Curious if others have explored the protocol layer or if most people are still building on top of the public API and wondering why results plateau.
The public Reddit API is rate limited, heavily monitored, and most importantly it identifies itself. When you make requests through the official API, Reddit's servers know exactly what you are. You're not a user. You're a script. That context is baked into how your traffic is scored and how your accounts are treated.
Reddit's actual iOS app doesn't use the public API at all. It runs on a completely separate internal client layer with its own authentication system, its own token structure, its own request signatures. The iOS app has a two tier token system that the public API doesn't replicate. It sends device attestation headers. It passes a device fingerprint with every request. All of this is invisible if you've only ever worked with the public API.
What this means practically: any tool built on top of the public Reddit API is operating at a disadvantage by design. You're using the front door Reddit built for developers to read data, not the same channel their own app uses to behave as a real user.
When I realised this I stopped building on the API entirely and started looking at what the iOS app was actually sending at the protocol level. Captured the full client session from a real iOS device, every identifier, every header, every token, and built infrastructure to replicate that at scale.
The result is something that behaves, from Reddit's perspective, exactly like a genuine iPhone running the Reddit app. Not an API call. Not a browser. An iPhone.
For anyone doing serious volume on Reddit this distinction matters more than almost any other variable. Curious if others have explored the protocol layer or if most people are still building on top of the public API and wondering why results plateau.