Riza has introduced new features to enhance the functionality and security of its Code Interpreter API, which facilitates the safe execution of untrusted code generated by large language models or humans. The API now accepts input files, allowing users to specify their paths for easier handling of multiple inputs, and it supports configuring credentials for outbound HTTP requests to securely authenticate without exposing sensitive data. Currently, the API supports HTTP Basic and Bearer authentication schemes, with plans to expand this in the future. Additionally, users have the option to run the Riza Code Interpreter API via a docker container for a self-hosted setup.