acheong08
Propose of this server
First of all, sorry for my ignorance.
I haven't clear the propose of this server, I was expecting a server API to can request directly to it the asking sentences in the same way the official webpage does to their backend, but I don't see any reference to chat.openai.com/backend-api endpoints in the code.
Looks like it is built to synchronize with an Agent somehow? Why then it is necessary? I'm missing many things.
Could you explain me more in detal how is it supposed to work? Always with an Agent? what's the role of this server? I would appreciate it so much :)
The issue:
- Cloudflare blocks can only be bypassed with a browser
- Rate limiting has increased
- OpenAI blocks IP addresses of VPS / hosted servers
My solution:
- Host API on VPS/server
- Run agents on local devices such as PC or Raspberry Pi via browser. This bypasses the IP blocks and cloudflare
- Multiple accounts can connect to a single API server. This gets around rate limits by cycling through connections
Further proposal:
- A public community API can be made where everyone can congregate their account resources and make a large connection pool. This will allow much faster response times and also higher stability that is not affected by a single connection error.
All interaction to ChatGPT's API is done on the agent side. The server is simply a remote endpoint that can pool connections
Using websockets, requests to the server can be instantly forwarded to the agents even if they're behind a NAT
