/gpt-server

gpt-server with gemini and groq as sever for Copilot like client

Primary LanguagePython

Use Gemini and Groq(Llama3, Mixtral) and OpenAI for Unary Interface

Easy Self Host setup for any usage, like CopilotChat.nvim

Make sure make is installed in your system, or your should run command in Makefile manually.

Add Support for Auto Switch between Gemini Pro 1.5 and Gemini Pro 1.0

Pre-Requirements

Make sure you have docker.

Change password in your Makefile for redis server

# change password
    docker run --name redis -p 6379:6379 -e REDIS_PASSWORD=123456 -d redis /bin/sh -c 'redis-server --appendonly yes --requirepass ${REDIS_PASSWORD}'

personal information file .person in root

groq=<YOUR API KEY>
gemini=<YOUR API KEY>
proxy=<YOUR PROXY>
redis=<REDIS URL>
redis_port=<REDIS Port>
redis_password=<REDIS PASSWORD>

Install Dependencies and Run Server

make run

Run in docker

make new

Rebuild docker

make rebuild

If you want to add new Llm servers.

  • copy and past file in routers
  • adjust some logic and add
  • add routers to __main__.py