Codegate

A configurable Generative AI gateway, protecting developers from the dangers of AI.

Features

• Secrets exflitration prevention
• Secure Coding recommendations
• Preventing AI from recommending deprecated and / or malicious libraries

Installation

Requirements

• Docker
• Docker Compose
• jq
• vscode

Steps

Download the installation script docker-compose.yml

Run the installation script

chmod +x install.sh && ./install.sh

The script will download the Continue VSCode extension, create a configuration file. The script will also create a docker-compose.yml file and start the services.

Usage

Tap the Continue button in the VSCode editor to start the service to bring up a chat window. The chat window will be displayed in the VSCode editor.

Usage

Basic Usage (Manual)

Start the server with default settings:

codegate serve

Custom Configuration

Start with custom settings:

codegate serve --port 8989 --host localhost --log-level DEBUG

Configuration File

Use a YAML configuration file:

codegate serve --config my_config.yaml

Example config.yaml:

port: 8989
host: "localhost"
log_level: "DEBUG"

Environment Variables

Configure using environment variables:

export CODEGATE_APP_PORT=8989
export CODEGATE_APP_HOST=localhost
export CODEGATE_APP_LOG_LEVEL=DEBUG
codegate serve

Development

Setup

# Clone the repository
git clone https://github.com/stacklok/codegate.git
cd codegate

# Create and activate a virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install development dependencies
pip install -e ".[dev]"

Testing

pytest

Running from image

A docker image can be built just with make image-build. That will start a codegate server ready to use. Then it can be started with:

docker run -p 8989:8989 codegate:latest

Additionally if you want to start with a pre-created database, a volume can be mounted:

docker run -p 8989:8989 -v /path/to/volume:/app/weaviate_data codegate:latest