Tech Stack

CLI

For the CLI development we decided to use Typer (https://typer.tiangolo.com/) since this is easy-to-use library on Python that allows to build CLIs fast.

Model

For the connection to the LLM model we decided to use LangChain (https://www.langchain.com/) since this is easy-to-use library for building AI agents.

Static analysis

Ruff

Ruff was chosen for its efficiency (written on Rust) and simplicity. It works as both linter and formatter.

Mypy

Mypy was chosen for its proven reliability in type check.

Testing

Pytest

Pytest was chosed for its easiness to use and fixtures functionality.

Analytics

We use Grafana since this instrument allows to easily do the analytics for any application.

Observability

We use OpenTelemetry since this is powerful tool that also is easy to install. Specifically, now we can see which methods take the time.