The author and their collaborator created a DIY Babel fish system that enables two people to communicate in any language, using Nexmo's Voice API and Microsoft Translator Speech API. The system consists of a Tornado web server that handles incoming calls, translates speech in real-time, and speaks the translated text back to the caller. The authors set up their environment, created a Nexmo application, and obtained keys for the Microsoft Translator Speech API. They then implemented the CallHandler, EventHandler, and WSHandler to handle the call flow, events, and WebSocket communication with the translator API. The system can be used for bi-directional calls, allowing users to communicate in their chosen languages.