November 2021 Summaries

National Native American Heritage Month is celebrated by exploring topics related to Native American history and contemporary life. Before European colonization, North America was home to diverse native societies with rich cultures, deep knowledge of their environments, complex belief systems, and long-standing traditions in music and visual arts. Initial contact between natives and colonists varied widely, but the years following were disastrous for native peoples due to disease, warfare, and population decline. US policies of removal and forced assimilation further impacted native communities. More recently, Native Americans have begun to shift their narrative through contributions to WWII and involvement in the Civil Rights movement. Language revitalization programs are crucial for preserving native languages and cultures. Today, Native Americans continue to innovate ways to preserve and share their heritage.

Luke Oliff has joined the Developer Relations team at Deepgram as a Senior Developer Experience Engineer. He brings nearly five years of experience in Developer Relations, emphasizing the importance of reducing friction and improving developer experiences with their products. Originally from Colchester, an ancient town rich in history, Luke is also a dedicated dad, active on Twitter, and enjoys playing PC games. In his streams, he covers topics like Nuxt, Vue, and Tailwind projects while donating stream income to charities such as BlackGirlsCODE and GirlsWhoCode. Deepgram encourages feedback from users through their GitHub discussions.

This tutorial provides a step-by-step guide on how to get started with Supabase, an open-source alternative to Firebase that offers Postgres as a service. It covers setting up a database for a Reading List app and demonstrates how to add, read, edit, and delete data from it using Supabase's API documentation. The author also briefly touches upon the authentication and storage services offered by Supabase but does not cover them in detail.

This article provides an introduction to JSON (JavaScript Object Notation), a lightweight text-based notation used for representing structured data. It covers the basics of JSON syntax, including key-value pairs and the use of double quotes around keys. The article also discusses JSON's agnostic nature, its support for various data types, and how to convert JSON to other languages like Node and Python. Additionally, it demonstrates how to extract specific data from a real JSON object using an example from Deepgram API response.

The uniqueness of one's voice is determined by the shape and size of their vocal tract, which coordinates muscles in the chest, throat, and mouth for articulation. Language change can be attributed to small changes in neural impulses sent to these muscles, influenced by social interactions and relationships. Dialects are a result of cooperation and shared experiences within communities, representing unique ways of speaking that evolved over time. Words get their fixed meanings based on the communicative needs of the community, with specialized communication being crucial for success in various fields such as sailing, physics, music, dance, war, etc. Businesses and academic disciplines also have their own dialects to facilitate effective communication among members. The genetic and memetic aspects contribute to an individual's voice, representing both their ancestry and aspirations.

This tutorial demonstrates how to create live speech transcriptions within a browser using the Deepgram API. The project is entirely browser-based and requires an API key, which should be kept secure as it provides access to Deepgram's APIs. The code sets up a WebSocket connection with Deepgram, sends audio data in 250ms increments, and handles the responses from Deepgram by displaying the transcriptions on the web page. The final code is provided for reference.

Michael Jolley has joined Deepgram as the Head of Developer Relations. With a background in polyglot development and a passion for helping others achieve "Ah-ha!" moments, he aims to create content, engage with communities, host hackathons, and provide resources to help developers level up their skills. Deepgram's developer relations team is focused on supporting developers in achieving success, regardless of whether they use Deepgram's APIs or not. Feedback and suggestions are welcome via GitHub discussions.

Deepgram has announced the final speaker lineup for its inaugural What's Next in Voice Summit, featuring executives from Volley, Red Box, and other leading companies discussing industry trends and cutting-edge use cases. The virtual event will take place on November 18th, highlighting emerging voice startups and emphasizing the importance of voice technology across various industries such as retail, finance, healthcare, and space. Topics covered during the summit include the resurgence of voice technology, voice being the next interface for all companies, using voice at NASA and beyond, and the future of AI and voice technology.

Deepgram has announced the release of five new language models and three new use case speech models, along with improvements to their Hindi and Spanish language models and Conversational AI and Meetings models. These enhancements aim to improve accuracy for transcription applications. The company's transfer learning approach and End-to-End Deep Learning architecture have enabled them to achieve high accuracy on these new speech models. Deepgram has also made its French and German language models free for the first 30 days of use, allowing users to collect more audio data for training and improvement.

This article provides a detailed guide on how to send audio files to an Express.js server using Node.js and Multer middleware package. It covers setting up the backend server with Express.js, implementing Multer for handling different file types, creating a POST endpoint to receive data from the client-side, and sending audio files from the front-end using HTML forms or Javascript FormData objects. The article also explains how to configure storage options in Multer and handle multiple files uploads.

Kevin Lewis, a Developer Advocate at Deepgram, shares his journey into the world of software development and developer relations. He started as an early-career developer after attending his first hackathon ten years ago. Since then, he has attended numerous hackathons, volunteered to help others, and eventually worked for the company that hosted his first event. Lewis's passion lies in supporting early-career developers by teaching them both technical skills and core skills necessary for a successful work life. In addition to his role at Deepgram, he runs an event series focused on these core skills. He enjoys playing board games and has a degree in Creative Computing. Kevin is active on Twitter (@_phzn) and encourages feedback through GitHub discussions.

This article provides a step-by-step guide on how to generate WebVTT and SRT captions automatically using Node.js with the help of Deepgram's API. The author explains the importance of captioning for accessibility purposes, especially in situations where audio might be difficult to hear or understand. The two formats discussed are very similar but have slight differences in their syntax. The article also introduces a new feature in Deepgram's Node.js SDK that simplifies the process of generating captions by providing .toWebVTT() and .toSRT() methods.

Sandra Rodgers has joined the Developer Relations team at Deepgram as a Developer Experience Engineer. She previously worked on the UXDev team at a fintech company and has experience in education. Rodgers is excited to help improve the developer experience for those in the Deepgram community.

This article provides a step-by-step guide on transcribing videos using Node.js and the Deepgram API. The process involves converting video files into audio files, which are then sent to Deepgram for transcription. The author guides readers through setting up necessary dependencies, creating utility functions for ffmpeg commands and file downloads, and writing functions for transcribing both local and remote videos. The complete project is available on GitHub for further reference.

The inaugural Deepgram Summit on November 18 aims to bring together thought leaders from various industries to discuss the future of voice technology and its potential applications across sectors such as retail, healthcare, finance, and space. With advancements in AI and NLP, modern speech recognition solutions can provide rich customer insights and power new applications, giving businesses a competitive advantage. The event will feature speakers from leading enterprises and high-growth startups who are leveraging the power of voice technology.

Brian Barrow has joined the Developer Relations Team at Deepgram as a Developer Experience Engineer. He started his career in Supply Chain Management and transitioned to programming after discovering his love for teaching and helping others learn new things. Brian is passionate about the opportunities programming provides, such as problem-solving and working with smart people. In addition to his work at Deepgram, he enjoys spending time with his family and playing disc golf.

FFmpeg is an open-source toolkit for converting and manipulating both audio and video files from the terminal. It is widely used for tasks such as file format conversion, trimming audio, changing volume, reducing background noise, and overlaying multiple audio files. The article provides a beginner's guide to using FFmpeg for audio-related tasks, covering basic commands, options for handling existing output files, minimizing terminal information, and working with simple and complex filters.

The article discusses a project called Bleepgram, which was developed during an internal hackathon at Deepgram. The purpose of the project is to automatically censor profanity in audio files using Node.js and Deepgram's Speech Recognition API. The process involves getting a transcript from Deepgram, identifying profane words, determining clean audio timings, creating complex FFmpeg filters for audio manipulation, and finally running FFmpeg to generate the censored file. The complete project code is available on GitHub at https://github.com/deepgram-devs/censor-audio.

In this article, Kevin Lewis demonstrates how to transcribe YouTube videos using Node.js and Deepgram's Speech Recognition API. The process involves downloading the video, converting it to an mp3 audio file, generating a transcript with Deepgram, storing the transcript in a text file, and deleting the media file. Key packages required for this project include @deepgram/sdk, ffmpeg-static, and youtube-mp3-downloader. The final project code can be found at https://github.com/deepgram-devs/youtube-transcripts.

This blog post provides an introduction to APIs (Application Programming Interfaces), their uses and how to make API calls using various programming languages such as JavaScript, Python, and Node.js. The author explains the basics of APIs, including their role in allowing developers to interact with platforms or services, and compares them to a restaurant menu where options are presented along with descriptions. The post covers some popular APIs like TheAudioDB API for music metadata, Google Maps Directions API, Stripe payment processing API, and Deepgram's AI Speech Recognition API. It also discusses the lego block analogy of APIs, where they provide small pieces of functionality that developers can build upon. The author then demonstrates how to make an API call using cURL in a terminal, explaining the two required components: endpoint (a URL) and method (GET, POST, PUT, DELETE). The post also covers URL parameters, data inclusion, headers, and when to use each component. Understanding HTTP status codes is crucial for working with APIs, as they indicate whether a request was successful or not. The author provides examples of common 400-level error codes related to API usage. The post concludes by showing how to make API requests in the browser using JavaScript's fetch() function and in Node.js using the node-fetch library. It also covers making API requests in Python using the requests library. The author provides a glossary of terms commonly used with APIs, such as rate limit, SDK, GraphQL, JWT, and OAuth. Overall, this post serves as an informative guide for developers new to working with APIs, providing practical examples and explanations of key concepts.

Deepgram has recently launched a new Developer Platform featuring updated API documentation and various tools for developers. The platform includes SDKs & Tools, a developer blog with content targeting different learning levels, and use cases with code samples to help users get started quickly. Deepgram is committed to helping developers build and learn, and encourages feedback on their GitHub discussions.

The article discusses sales and support enablement development paths, focusing on speech recognition technology that enables situational training, analysis, and coaching for sales and support teams. It outlines three main use cases - onboarding, coaching, and real-time assistance - and provides a high-level process flow of the various processes involved in these software solutions. The article also offers tips for choosing a development path, emphasizing the need to identify and prioritize key needs, fix issues with appropriate resources, measure success, and create win-win situations with solution providers.