Glide is a cloud-native LLM gateway, delivering high-performance LLMOps in a lightweight package.
Glide provides a simplified way to build application resilience, reduce latency, and manage API keys among many other features.
A single model setup is not recommended for production deployments.
The initial phase in deploying Glide involves setting up your configuration file in the project root. For the basic setup, it is essential to have at least one pool with a minimum of one model.
Install Git:
Ensure that Git is installed on your machine. You can download it from Gitβs official website.
Clone the Repository:
Open a terminal or command prompt and paste the code below:
Copy
git clone https://github.com/EinStack/glide.git
Build Docker Image:
Create a Dockerfile in the project root. Below is a sample.
Copy
# syntax=docker/dockerfile:1FROM golang:1.22rc1-alpine3.19# Set destination for COPYWORKDIR /app# Download Go modulesCOPY go.mod go.sum ./RUN go mod download# Copy the source code. Note the slash at the end, as explained in# https://docs.docker.com/engine/reference/builder/#copyCOPY . ./# BuildRUN go build -o ./dist/glide# Update HostPort in pkg/api/http/config.go to 0.0.0.0:9099. Eventually this will be specified in the config file.EXPOSE 9099 # Run # replace with the name of the config file in rootCMD ["./dist/glide", "-c", "./config.yaml"]
Now that you have a Dockerfile, build the container.
Copy
docker build -t <image_name> .
Deploy:
Deploy the image to your prefered image hosting platform.