March 2021 Summaries

In a detailed walkthrough, the process of using Postman and AssemblyAI for speech-to-text transcription without writing any code is explained. By leveraging Postman, a versatile API testing tool, users can upload local audio files to AssemblyAI, which is an API-based transcription service, and receive a URL for the uploaded file. The guide outlines how to configure HTTP requests in Postman to first upload an audio file and then request its transcription by using the URL provided by AssemblyAI. The transcription process includes the option to enable advanced features such as speaker labels and PII redaction. Once the transcription is processed, users can retrieve it by making a GET request with the transcription ID. This method provides a quick and efficient way to utilize AssemblyAI's transcription services, demonstrating how API collections can be shared among teams depending on the Postman licensing tier.

The blog post expands on a previous tutorial by demonstrating how to enhance a simple command-line application with file-upload functionality using Node.js and the AssemblyAI speech-to-text API. It guides users through the process of uploading an audio file directly from a local system to AssemblyAI by setting up a Node.js environment, creating a new JavaScript file, and using Node.js libraries such as `fs` for file streaming and `node-fetch` for HTTP requests. The tutorial provides detailed instructions for reading an audio file into memory, sending it as chunked data to AssemblyAI's API, and handling the response, which includes a URL to the uploaded file. This URL can then be used in further stages of the application to retrieve the transcription ID and ultimately the transcribed text. The post encourages users to follow along with the code available in a GitHub repository and offers troubleshooting tips for potential errors during the process.

This tutorial guides you through creating a basic Node.js command-line interface (CLI) app that uses the AssemblyAI speech-to-text transcription API. The application first uploads an audio file URL to the API, then fetches and prints the completed transcription text on the command line. To get started, you need a free AssemblyAI account, a code editor like VS Code, Node.js, and npm installed. You will create three new files: upload.js, download.js, and .env. The upload.js file sends an HTTP POST request to the AssemblyAI API with the audio URL as JSON data. The download.js file checks if the transcription is completed by sending a GET request to the AssemblyAI endpoint for retrieving transcriptions. If the status is "completed," it prints the text on the command line; otherwise, it prompts you to try again in a few minutes.

AssemblyAI has significantly improved its speech-to-text features, including punctuation and casing restoration. The company's new model is a multi-class classifier that predicts actions such as adding punctuation or changing casing for each word in the transcription. This transformer-based model architecture yields an accuracy of over 92% for punctuation and casing restoration, trained on over 1 billion tokens. The model performs exceptionally well even with industry-specific language. Punctuation and casing are applied by default to all API requests, making it easy for users to utilize this new feature.