Self-host llmapi web chat for ChatGPT and others
llmapi-chat is a simple web-chat-server
backend that can chat with a variety of large language models (LLM, such as ChatGPT, GPT-3, GPT-4, etc.)
- Continuous Dialogue: Use
ChatGPT
for multiple rounds of continuous dialogue - Text completion: use
GPT-3
orwelm
model for text completion - Text embedding: use openai's
Embedding
model to convert sentences into vectors, which can be used for classification, search and other tasks - Text to image: use
dall-e
model to convert text to image - Real-time search: Use
NewBing
interface to experience Bing search (based on ChatGPT, unofficial interface)
- build and run locally
./build.sh
./start.sh
- run with docker
docker build -t llmapi-chat:0.1.0 .
docker run --name webchat -d -p 8081:8081 llmapi-chat:0.1.0
Open
http://127.0.0.1:8081
in the browser to visit the webchat page.
You can sign up on llmapi.io to get apikey
, setup as below to chat with LLMs like ChatGPT
GPT3
and so on.
This project developed based on chatGPT-web