Latest insights in
Human Computing

Virtual humans: the business case for AI video agents that see, hear, act, and respond

AI video agents that see, hear, and respond in real-time conversations. The enterprise business case for virtual humans across healthcare, L&D, insurance, and recruiting.

AI faces: how they're generated, animated, and used

Modern AI faces can hold live conversations and read your expressions. Explore the rendering, turn-taking, and perception models that make it possible. 

Face-to-face conversational AI vs. chatbot: understanding the full spectrum

The face-to-face conversational AI vs. chatbot comparison comes down to capability tiers. A guide to types, examples, and when to move from text to video.

Customer service training: AI video agents for realistic practice

Discover how AI video agents transform customer service training with realistic practice. Learn implementation strategies, benefits, and real-world applications.

Conversational AI for healthcare: Building HIPAA-compliant patient video conversations

Conversational AI for healthcare is evolving beyond chatbots. Learn how to build HIPAA-compliant patient video conversations with real-time AI Personas.

Learnings From Random LoRA-Finetuning Experiments On A Guided Music Synthesis Diffusion Model

A hands-on experiment fine-tuning ACE-Step v1.5, an open-source AI music generation model, to fix its garbled lyric problem using LoRA and Attention Sharpness Regularization. What started as a random Friday night side project turned into a deep dive into how a 2-billion-parameter diffusion transformer actually processes text, and what breaks when you try to change it.

Phoenix-4: Real-Time Human Rendering with Emotional Intelligence

Phoenix-4 is the first real-time model to generate and control emotional states, active listening behavior, and continuous facial motion as a single, unified system. It is a real-time behavior generation engine, built from the ground up, that goes beyond photorealism to transform conversation data into emotionally responsive, context-aware facial expression and head motion with millisecond-level latency.

Raven-1: Bringing Emotional Intelligence to Artificial Intelligence

Introducing Raven-1. A multimodal perception system that captures not just what users say, but how they say it, how they look when they say it, and what that combination actually means. It interprets tone, expression, hesitation, and context in real time, enabling AI that can truly understand intent rather than simply respond to words.

Tavus flight log: 2/6 + welcome :)

We’re building a lot and shipping faster than ever, and with that, I’m going to start writing a weekly update each Friday for a small group of our early adopters and power builders. My goal is to give a raw view into what we’re doing over here as we build together to bring human computing to life.

Sparrow-1: Human-Level Conversational Timing in Real-Time Voice

Sparrow-1 is a specialized, multilingual audio model for real-time conversational flow and floor transfer. It predicts when a system should listen, wait, or speak, enabling response timing that mirrors human conversation rather than simply responding as fast as possible.

Santa is real. And he’s back as a PAL

Last year, millions of people talked to AI Santa through Tavus. That is why we can say this with confidence: Santa is real. And now he is back.

From random noise to real images: Understanding loss-function formulation

From theory to code: how diffusion and flow matching models turn Gaussian probability paths into vector fields, training targets and practical loss functions.

How Tavus brings human computing to life

Discover how Tavus human computing transforms AI into lifelike, engaging conversations that boost retention and speed—experience the future of interaction today.

Using Tavus to your unfair advantage: AI workplace consultant

A practical guide to using Tavus to build a hyper-contextual AI workplace consultant that helps you make better decisions, get promoted, and stay sane.

Work Trial AI is fixing the broken job market

Work Trial AI is redefining hiring by replacing outdated interviews with real-world, AI-powered work trials. Built by Final Round AI, it helps candidates showcase skills authentically and lets companies see how people actually work.

The targets you can’t see

Learn how good intentions turn into empty metrics and what it takes to stay true to your goals. This post breaks down the steps teams take away from truth and how to bring real meaning back to your strategy.

How to build an AI recruiter that gives every applicant a fair shot

AI recruiters let every applicant in your ATS speak, gather richer insights through lifelike video interviews, and deliver structured reports that streamline hiring.

From random noise to real images: Understanding diffusion and flow matching

A clear intro to diffusion and flow-matching: data distributions, ODE vs SDE, and the path from Gaussian noise to realistic images/videos powering SOTA models.

Tech meets the environment: How Brian Johnson is tackling trash

How Brian Johnson, Tavus technologist, built a drone and AI-powered system to detect and report illegal trash dumping. With 700+ flights, 100k images, and 6k+ sites reported, his project uses tech to deliver environmental impact.

What is a prompt injection attack?

Prompt injection is the new social engineering for AI—tricking models into ignoring safeguards, leaking data, or running harmful actions. Learn how attacks work, why OWASP ranks them the top AI risk, and what companies like Anthropic are doing to defend against them.

How AI is affecting the job market

About 40% of jobs are exposed, adoption is rising fast, and AI skills carry a 56% wage premium. Here’s what the data shows and what to do in the next 90 days.

Four quickstart use cases for Tavus

Discover four quickstart use cases with Tavus—including sales, onboarding, recruiting, and healthcare—that you can ship quickly and start seeing results.

Introducing Persona Builder: AI personas that feel uniquely yours

See how the Tavus Persona Builder empowers you to design custom AI humans with unique personalities, voices, and expertise tailored to your audience.

Introducing Knowledge Base: AI personas that know their stuff

Explore how Tavus’s Knowledge Base equips AI humans with your company’s information so every conversation is accurate, consistent, and on brand.

How iAsk brings AI tutors to students across the globe

Discover how iAsk uses Tavus Conversational Video Interface to power on-demand AI tutors for 22,000+ students each month—delivering fast, reliable answers that Gen Z trusts.

Introducing Objectives & Guardrails: AI personas that stay on track

Learn how Objectives & Guardrails in Tavus keep AI humans on track, delivering conversations that are reliable, brand-safe, and aligned with your goals.

Introducing Memories: AI that actually remembers

Discover how Tavus Memories lets your personas retain context across conversations, creating natural, continuous interactions that feel genuinely human.

Tavus brings human computing to Ai4

At Ai4, Tavus drew nonstop traffic, introduced thousands to human computing, and proved AI Personas are the category-defining future of enterprise AI.

Why I joined Tavus as the Head of Growth Marketing

In June, I joined Tavus as the Head of Growth Marketing. Tavus's future is going to be monumental—here's why I'm excited and humbled to be a part of it.

How Final Round AI scales realistic mock interviews with Tavus CVI

How Final Round AI uses Tavus CVI to deliver lifelike mock interviews at scale, helping 100,000+ users land jobs faster with real-time AI.

How to build a conversational video AI with a transparent background

Learn how to build a real-time conversational video AI with a transparent background using Tavus CVI, WebGL, and Daily WebRTC. Perfect for overlays, product demos, and coaching apps—deployable in minutes.

Conversational video AI cost comparison

When you’re evaluating tech as new as conversational video AI, pricing isn’t just a line item—it shapes how you experiment, scale, and ultimately succeed.

Smarter, faster, fairer: How AI is reshaping the future of recruiting

A recent global survey from Korn Ferry found that 67 % of talent acquisition professionals see increased AI usage as one of the top trends for 2025.

How creating Sparrow made me a better conversationalist

Building Sparrow taught me that timing is everything. Here's how AI turn-taking made me a better engineer—and a better conversationalist.

Understanding intuition behind multi-turn LLMs through the prism of search

Discover the latest research in how LLMs use reinforcement learning to search, reason, and refine answers across multiple turns—boosting accuracy and enabling active problem-solving.

How to build a chatbot using conversational video AI in React

Building a video AI chatbot is more than just hooking up a webcam. It’s about orchestrating a set of sophisticated systems to deliver smooth, natural conversations.

How to deploy a Tavus bot as an internal sales/BDR coach

Step-by-step instructions on how to deploy a Tavus AI coach to scale personalized video feedback for sales reps. Automate training, improve pitches, and boost conversions—powered by your data.

The (Tavus) Hackathon Cookbook

So you want to build a Conversational Video Interface? You’re in good company, startups like Mercor and Delphi and fortune 500s alike are leveraging Tavus APIs to build the future of AI interaction.

Hummingbird-0: Advancing Zero-Shot Lip Synchronization in AI-Generated Video

We made an unexpected discovery while developing our premium conversational AI technology. Components of our advanced video pipeline could be isolated and explicitly optimized for lip synchronization, with remarkable results. This serendipitous research byproduct evolved into Hummingbird, a specialized zero-shot lip-sync model that achieves state-of-the-art performance compared to other leading solutions.

Sparrow-0: Advancing Conversational Responsiveness in Video Agents with Transformer-Based Turn-Taking

In this paper, we dive into the development and research behind Sparrow-0, exploring the innovative transformer-based approach for turn-taking and its integration alongside Raven and Phoenix models within our Conversational Video Interface (CVI), an end-to-end operating system designed for building responsive video agents.

Introducing the evolution of Conversational Video Interface – now with Emotional Intelligence

Introducing our new family of state-of-the-art AI models: Phoenix-3, Raven-0, and Sparrow-0. Together they bring Conversational Video Interfaces (CVI) to the next level, and power Charlie, our new demo persona.

Introducing AI Santa: The Magic of Christmas, Powered by Tavus

What if Santa could talk to everyone all at once? Now He Can! We are thrilled to announce AI Santa by Tavus, the world’s first fully AI-powered replica of the man in the red suit.

11+ Best Text to Speech APIs [2024]

Unlock the power of speech with our top picks for the best Text-to-Speech APIs of 2024! Find the perfect voice solution for your app today.

Open-Sourcing AI Innovation: Building Real-Time AI Interactions with Pipecat and Tavus

Pipecat + Tavus empowers developers to build modular, real-time conversational AI systems with low latency, vendor neutrality, and support for 40+ LLMs, STT, and TTS services.

VEED Leaps Ahead in AI Avatar Development by Leveraging Tavus APIs

What is an AI Tool? How Artificial Intelligence Tools Work [2025 Guide]

Discover the power of AI tools and understand how they revolutionize tasks with our comprehensive 2025 guide to artificial intelligence tools.

What Makes Conversational AI Human Like?

The Tavus team leveraged multiple techniques to get life-like conversational interface with an AI. In particular, streaming, parallelism, resource management, and memory management. This combination helped reduce a long utterance to utterance time down to under a second.

Tavus is SOC 2 Certified

Build a Custom Personality for Real Time Video AI

Customize your conversational AI's personality using a custom persona with system prompts, context, and your own LLM.

Build Your First Real Time Conversational Digital Twin in Five Minutes

A tutorial on how you can build a conversational AI in just a few minutes with details of what each of the parameters do.

Introducing: The world's fastest Conversational Video Interface for developers

Humanize digital interactions with real-time interactive digital twins that can speak, see, and hear.

How to Make AI Videos (Step-by-Step Guide for 2024)

Deploy AI video creation into your apps in 2025! Follow our easy step-by-step guide to integrate and produce stunning AI-generated videos that captivate audiences.

18+ Best AI Marketing Tools & Software [2024]

Discover the top AI marketing tools for 2024! Boost your campaign efficiency, drive sales, and outperform your competition today.

Phoenix-2: Advanced Techniques in Talking Head Generation — 3D Gaussian Splatting

This paper will cover the past, present and future of the talking-head generation research field. Specifically, we will dive deep into the trending 3D scene representations (NeRF -> 3DGS) and the benefits of employing 3DGS in avatar applications.

How to do Text to Video for AI Replicas

An introduction to how to make an AI video with Tavus' video API.

16+ Best AI Voice Generators [2025]

Discover the best AI voice generators for your applications to enable users to create custom videos, voiceovers, and even music.

20+ Best AI Avatar Generators (Video & Image) [2025]

Discover the best AI avatar generators in 2025 for integrating personalized avatars into apps and elevating user interactions.

Phoenix-1: Realistic Avatar Generation in the Wild

This research paper, written by the Tavus team, details the development of Phoenix, a groundbreaking generative model for realistic avatar creation and text-to-video generation. Phoenix leverages audio and text-driven 3D models, integrating volumetric rendering techniques and 2D Generative Adversarial Networks (GANs) to create lifelike replicas from short video clips.

Introducing Phoenix & the Tavus Developer Platform Public Beta

Users generate realistic videos from just a script with their own digital replica

11+ Best AI Video Editing Software

Learn about the best options available for AI video editing software and how developers can integrate this into their applications.

Character AI API Review & Alternatives [2025]

Explore Character AI's API capabilities and discover alternatives for 2025. Understand its strengths, limitations, and find the best API for your needs.

18+ Best AI Software [2024]

Discover the best AI software for 2024 in niches like AI video creation, imaging, sales, workflow automation, writing, data analysis, and chatbots…plus more!