March 2023 Summaries

ngrok-rs is a newly announced open-source crate designed to integrate secure ingress directly into Rust applications, simplifying the process of making Rust services accessible over the internet without requiring low-level network setups like IPs or certificates. By acting as a Rust-native version of the ngrok agent, ngrok-rs allows applications to listen on ngrok's global ingress network and seamlessly integrate with frameworks like hyper and axum. It eliminates the need for traditional sidecar agents, offering a high-level abstraction for managing network configurations across various environments, thereby enhancing application portability. ngrok-rs also supports integration with the tracing framework for logging and utilizes Rust’s FFI capabilities for building SDKs in other programming languages. The crate is part of a broader effort to create language-native SDKs, with ongoing development for languages like JavaScript and Python, and plans to expand to others, including C# and Ruby.

The 1Password ngrok shell plugin offers a streamlined solution for managing ngrok credentials by storing them in 1Password instead of a local ngrok.yml file, thus enhancing security and convenience. This integration allows users to authenticate via their 1Password vault and seamlessly access their credentials without manual updates, making it easier to work across different devices and environments. The setup involves installing the 1Password app and CLI, connecting them, and initializing the ngrok plugin. Users can choose how the plugin manages credentials, ranging from prompting for each session to using a directory-specific or global default approach. This tool not only simplifies credential management for individual users but also benefits teams by enabling centralized credential issuance and management, eliminating the need to log into the ngrok Dashboard. The author, initially skeptical, found the plugin easy to set up and beneficial in solving credential management challenges, particularly in maintaining synchronized credentials across devices.

ngrok has added security features to its free plan, including OAuth and Webhook validation, allowing developers to integrate up to 5 users a month with an OAuth provider and secure webhook communications up to 500 requests a month. The addition of these features is made possible by ngrok's integration with popular OAuth providers such as Google, Microsoft, and GitHub, which replaces the burden of implementing OAuth on the developer's part. Webhook validation is also simplified through ngrok's partnerships with 60 popular services, allowing developers to secure their webhook communications with a single command. The move aims to empower developers while keeping security teams happy, aligning with ngrok's mission to take ingress off developers' plates and make security-focused features accessible and easy to use.

ngrok has released ngrok-go, an open-source Go package that allows developers to embed secure ingress directly into their Go applications, simplifying the process of exposing Go apps to the internet. This new package modernizes the API by introducing features like Traffic Policy and offers an interface that is user-friendly and easy to maintain. ngrok-go allows applications to use ngrok's global ingress network with minimal configuration, eliminating the need for setting up complex network components like IPs, certificates, and load balancers. By calling ngrok.Listen(), developers can serve their applications online without managing low-level networking details, thus making the application portable across different environments such as VMs, Kubernetes, or even local machines. The package is designed to integrate seamlessly with existing Go network code, using functional options for configuration and logging adapters for popular Go libraries. ngrok also plans to expand support to other programming languages, with libraries for Rust and JavaScript already in development.

The blog post provides a detailed guide on building an SMS chatbot similar to ChatGPT using the OpenAI API and Twilio Programmable Messaging with Python. It begins with an introduction to ChatGPT, a conversational AI developed by OpenAI, and its capabilities in natural language processing. The tutorial outlines necessary prerequisites, including obtaining a Twilio account and phone number, setting up an OpenAI account, and installing Python and a few packages like Flask and python-dotenv. It explains the steps to create a project directory and virtual environment, install required Python packages, and configure environment variables, notably the OpenAI API key. The post walks readers through writing a Python script using Flask to create a web application that interacts with Twilio's messaging service, allowing users to send SMS messages that the OpenAI model processes and responds to. The application utilizes the text-davinci-003 model from OpenAI for text generation, similar to that used by ChatGPT, and includes instructions for running the app locally with ngrok to expose it to Twilio's servers. Finally, the post emphasizes the broad development possibilities enabled by combining ChatGPT with Twilio, encouraging readers to explore various applications.