A first of its kind experimental title, where your voice is the controller. Chat with quirky robots, help solve their problems, and shape the story through real-time conversations in this cozy sci-fi ...

Join Deepgram and Pipecat live to build voice agents with Flux TTS on Pipecat: consistent latency, interruption feedback, and reference code you can build on.

Deepgram self-hosted now runs in a FIPS 140-3-compliant configuration, built on a CMVP-validated cryptographic module. Run the best speech models on your own infrastructure and clear the federal procu...

This blog post explores the essential APIs needed to build effective programmable voice agents, detailing their functionalities, trade-offs, and best use cases.

This tweet promotes the blog post by highlighting the significant time sales reps spend on non-selling activities and introduces conversation intelligence as a solution to improve efficiency and CRM accuracy.

This blog post explores the unique audio challenges faced by restaurants and how Deepgram's technology is designed to enhance voice recognition in noisy environments.

This tweet promotes the importance of accurate speaker labels in speech-to-text transcription, encouraging readers to explore practical frameworks and insights discussed in the blog post.

Deepgram's new Australia endpoint is now available! This allows organizations with data residency needs to process audio securely within Australia while maintaining performance and compliance.

This tweet promotes the blog post about building live dashboards for real-time speech analytics, highlighting the importance of understanding latency and cost factors.

This tweet promotes the importance of understanding the difference between demo functionality and real-world production readiness in voice agent deployments, encouraging readers to learn more.

This tweet promotes the blog post on AI vs. legacy services in medical transcription, highlighting a key finding from a University of Wisconsin trial that shows significant time savings for clinicians.

Introducing Batch Diarization V2, our latest upgrade that enhances speaker labeling accuracy, crucial for effective data analysis in various fields. Learn more about its benefits!

This tweet promotes a blog post about running AI call center voice agents in production, highlighting orchestration challenges and key metrics for success.

This tweet promotes the blog post about how enterprise restaurant brands are leveraging advanced voice AI models, emphasizing the unique challenges and solutions provided by Deepgram.

Explore the differences between bundled and assembled voice agent APIs, how pricing influences architecture, and key considerations for your voice solutions in our latest blog post!

This tweet promotes the blog post on integrating voice agents with Salesforce, emphasizing the importance of choosing the right speech layer to avoid transcription errors that can compromise CRM data quality.

This tweet emphasizes the importance of ASR accent coverage in restaurant voice AI, highlighting how a single misheard word can lead to order errors, encouraging readers to learn more.

This tweet promotes the blog post about HIPAA business associate amendments for voice AI, emphasizing the importance of compliance with upcoming regulations.

Deepgram and Fortanix partner to deliver secure voice AI solutions for regulated industries, enabling enterprises to run voice AI in their own environments while protecting sensitive data.

This blog post discusses the differences between Cascade and Speech-to-Speech voice agent architectures, emphasizing their impact on costs, debugging, and compliance in production environments.

This tweet promotes the blog post on dynamic range compression (DRC) in voice AI, highlighting its potential benefits and urging readers to learn about the common misconceptions surrounding it.

This tweet promotes the blog post on AI voice agents in healthcare, highlighting the importance of the speech recognition model and its impact on reducing human intervention in calls.

This tweet promotes the blog post on the rapidly growing voice AI market for restaurants, highlighting the emergence of new companies and the importance of understanding the ecosystem.

This tweet promotes the blog post about Deepgram's Voice Agent API, highlighting its focus on data security and fast performance, encouraging readers to learn more.

This tweet promotes the blog post discussing the inadequacies of word error rate (WER) in evaluating speech recognition for Indian languages, introducing the BRIDGE 7-metric framework as a solution.

Explore the evidence behind AI voice agents in patient engagement. Learn about the potential benefits and the importance of validating vendor claims before implementation.

The blog post explains why many ASR models fail with Hinglish, the prevalent code-switched language in India, and offers solutions to improve accuracy for better user interactions.

This blog post provides essential insights on how Voice AI can impact compliance in healthcare, highlighting the importance of accuracy in transcription and risk assessment.

This tweet promotes a blog post discussing the challenges of code-switching in automatic speech recognition (ASR) systems, highlighting significant error rates and encouraging readers to learn more.

The Browser Agent SDK is four composable npm packages that drop a Deepgram voice agent into any web app, with a clean path from one-line widget to full framework-agnostic control.

A comprehensive comparison of Deepgram vs Rev AI, focusing on their strengths for real-time streaming or batch transcription. Key differences in deployment, pricing, and accuracy are explored.

This tweet promotes the blog post on call center compliance regulations for 2026, highlighting the importance of understanding and implementing the new AI-specific regulations.

This tweet promotes the blog post on chatbot vs. conversational AI, highlighting a critical statistic about self-service resolution rates to encourage readers to learn more.

This blog post explores the discrepancies between TTS playground demos and production APIs, highlighting the importance of real-world testing for accurate pronunciation.

This tweet promotes our blog post on medical voice recognition, highlighting the significant time physicians waste on documentation and the financial impact of burnout. It encourages readers to explore how to effectively evaluate voice recognition tools in healthcare.

Deepgram's Nova-3 enhances speech-to-text support across Asia-Pacific, introducing new languages and improved accuracy for existing ones, making voice AI more accessible and reliable.