Showing the first 18000 characters. Load the complete file to read the remainder.
# Tavus
- [Introduction](https://docs.tavus.io/sections/introduction.md): Use Tavus tools and guides to build PALs with real-time human-like perception and presence, bringing the human layer to AI.
- [Agents & Automation](https://docs.tavus.io/sections/agents-and-automation.md): How to access machine-readable docs and API artifacts for developers, IDEs, and automation - llms.txt, OpenAPI, Agent Skills, and MCP.
- [What Is a PAL?](https://docs.tavus.io/sections/conversational-video-interface/pal/what-is-a-pal.md): A short definition of PAL, how it relates to Face and Conversation, and where to configure one.
- [What Is CVI?](https://docs.tavus.io/sections/conversational-video-interface/overview-cvi.md): CVI enables real-time, human-like video interactions through configurable PALs.
- [Sales Development Representative](https://docs.tavus.io/sections/conversational-video-interface/conversation/usecases/sales-development-representative.md): Qualify leads on a live call and sync results to your CRM.
- [AI Interviewer](https://docs.tavus.io/sections/conversational-video-interface/conversation/usecases/ai-interviewer.md): Engage with the AI Interviewer PAL to run structured, conversational interview simulations.
- [Customer Service Agent](https://docs.tavus.io/sections/conversational-video-interface/conversation/usecases/customer-service.md): Engage in real-time customer support conversations that adapt to user emotions and behavior.
- [Medical Intake](https://docs.tavus.io/sections/conversational-video-interface/conversation/usecases/medical-intake.md): Pre-visit intake flow with structured data capture.
- [Models](https://docs.tavus.io/sections/models.md)
- [PAL Strategies](https://docs.tavus.io/sections/onboarding-guide/pal-strategies.md): Two approaches to using PALs at scale: reuse with conversational context vs. create-per-session. Choose based on whether you need different data or different behavior per conversation.
- [Tavus Tool Calling](https://docs.tavus.io/sections/onboarding-guide/tool-calling-examples.md): Key guidelines for implementing reliable tool calls in Tavus conversational agents
- [Reducing Join Latency](https://docs.tavus.io/sections/onboarding-guide/latency-optimization.md): Strategies to reduce perceived and actual latency when starting Tavus conversations
- [Prompting Guide](https://docs.tavus.io/sections/onboarding-guide/prompting-guide.md): The recommended structure for writing system prompts for Tavus CVI PALs - identity, style, behaviors, and guardrails that work in real-time video conversations.
- [Objectives Prompting Guide](https://docs.tavus.io/sections/onboarding-guide/objectives.md): Write clear, testable prompts for objectives with practical examples.
- [EU AI Act Compliance](https://docs.tavus.io/sections/onboarding-guide/eu-ai-act.md): Product controls for biometric emotion recognition and AI disclosure, relevant to EU AI Act configuration of PALs and conversations.
- [API Conversation Quickstart](https://docs.tavus.io/sections/conversational-video-interface/quickstart/cvi-quickstart.md): Create your first PAL using the full pipeline and start a conversation in seconds.
- [Overview](https://docs.tavus.io/sections/agent-tools/overview.md): Build a PAL, wire its tool calls into your own code, and test the integration from your editor with the MCP server or your terminal with the CLI.
- [MCP Server](https://docs.tavus.io/sections/agent-tools/mcp-server.md): Connect your coding agent to the hosted Tavus MCP server to build PALs, wire bidirectional tool calls into your own code, and test the integration end to end.
- [MCP Tools Reference](https://docs.tavus.io/sections/agent-tools/mcp-tools-reference.md): Complete reference for every tool exposed by the Tavus MCP server, grouped by category.
- [CLI](https://docs.tavus.io/sections/agent-tools/cli.md): Install, authenticate, and drive the Tavus CLI to manage PALs, faces, conversations, and the agentic resource library from your terminal.
- [Agentic PAL Building & Testing](https://docs.tavus.io/sections/agent-tools/pal-build-and-verify.md): Use the Tavus MCP server or CLI to build a new PAL in a loop: adjust configuration, run simulated CVI chat turns, judge the result, and refine until the PAL behaves as intended.
- [Pipeline Modes](https://docs.tavus.io/sections/conversational-video-interface/quickstart/pipeline-modes.md): Run CVI with the full Tavus pipeline, Echo mode, or integrate via LiveKit, Pipecat, or Agora.
- [Overview](https://docs.tavus.io/sections/conversational-video-interface/pal/overview.md): Define how your PAL behaves, responds, and speaks by configuring layers and pipeline modes.
- [Stock PALs](https://docs.tavus.io/sections/conversational-video-interface/pal/stock-pals.md): Tavus offers pre-built PAL templates to help you get started quickly.
- [PAL Editing Lifecycle](https://docs.tavus.io/sections/conversational-video-interface/pal/editing-lifecycle.md): Once a PAL has a deployment, PAL Maker edits become drafts you push. API edits keep going straight to the live PAL.
- [Objectives](https://docs.tavus.io/sections/conversational-video-interface/pal/objectives.md): Objectives are goal-oriented instructions to define the desired outcomes and flow of your conversations.
- [Guardrails](https://docs.tavus.io/sections/conversational-video-interface/guardrails.md): Guardrails provide your PAL with strict behavioral guidelines that will be rigorously followed throughout every conversation.
- [Knowledge Base](https://docs.tavus.io/sections/conversational-video-interface/knowledge-base.md): Upload documents to your knowledge base for PALs to reference during conversations.
- [Memories](https://docs.tavus.io/sections/conversational-video-interface/memories.md): Help a PAL remember each participant and build continuity across conversations.
- [Overview](https://docs.tavus.io/sections/faces/overview.md): Understand Tavus Faces, learn how to create your own using a video or an image, or explore ready-to-use stock Faces.
- [Video to Face](https://docs.tavus.io/sections/faces/video-to-face-quickstart.md): Create a Phoenix face from a training video using the Create Face and Get Face APIs.
- [Image to Face](https://docs.tavus.io/sections/faces/image-to-face-quickstart.md): Create a Phoenix face from a headshot image using the Create Face and Get Face APIs.
- [Phoenix-4.5 Preview and Status](https://docs.tavus.io/sections/faces/phoenix-45-preview-and-status.md): How status and finetune_status work while a Phoenix-4.5 face goes from zero-shot preview to tuned.
- [Voices for Image-Based Faces](https://docs.tavus.io/sections/faces/voices-for-image-based-faces.md): Pick a voice when you train a face from a photo: use your own, or pick a stock voice to get running faster.
- [Choose a Model and Training Path](https://docs.tavus.io/sections/faces/which-training-path.md): Pick a Phoenix model, then decide whether to train from a photo or a video.
- [Phoenix-4.5 Image Requirements](https://docs.tavus.io/sections/faces/phoenix-45-image-requirements.md): Photo rules for training a Phoenix-4.5 face from an image.
- [Phoenix-4.5 Video Requirements](https://docs.tavus.io/sections/faces/phoenix-45-video-requirements.md): Recording rules for training a Phoenix-4.5 face from video.
- [Phoenix-4 Image Requirements](https://docs.tavus.io/sections/faces/phoenix-4-image-requirements.md): Photo rules for training a Phoenix-4 face from an image.
- [Phoenix-4 Video Requirements](https://docs.tavus.io/sections/faces/phoenix-4-video-requirements.md): Recording rules for training a Phoenix-4 face from video, including full body.
- [Phoenix-3 Video Requirements (Legacy)](https://docs.tavus.io/sections/faces/phoenix-3-video-requirements.md): Legacy recording rules for training a Phoenix-3 face from video.
- [Emotion Control with Phoenix-4 and Phoenix-4.5](https://docs.tavus.io/sections/conversational-video-interface/quickstart/emotional-expression.md): Unlock emotionally expressive facial movements and micro-expressions using Phoenix-4 and Phoenix-4.5 faces.
- [Stock Faces](https://docs.tavus.io/sections/faces/stock-faces.md): Start a conversation instantly with our ready-to-use stock faces - no training required. Stock faces are prebuilt, high-quality faces designed for natural, expressive communication. They enable lifelike interactions with dynamic facial expressions, smooth turn-taking, and consistent, high-quality vi…
- [Stock Face Model Map](https://docs.tavus.io/sections/faces/stock-face-model-map.md): UUID map so you can track Tavus stock faces and their looks across multiple Phoenix models. Always use the latest when available!
- [Face FAQs](https://docs.tavus.io/sections/faces/face-faqs.md)
- [Voices](https://docs.tavus.io/sections/conversational-video-interface/voices.md): Create a voice once, then use it by ID on any number of PALs and Faces.
- [Text-to-Speech (TTS)](https://docs.tavus.io/sections/conversational-video-interface/pal/tts.md): Configure how your PAL sounds. Set a Tavus Voice by ID, or bring a voice from your own TTS provider account.
- [Language Support](https://docs.tavus.io/sections/conversational-video-interface/language-support.md): Declare the languages a conversation may run in, and see which languages Tavus can speak and hear.
- [Pronunciation Dictionaries](https://docs.tavus.io/sections/conversational-video-interface/pal/pronunciation-dictionaries.md): Control how your PAL pronounces specific words, names, and terms during conversations.
- [Conversational Flow](https://docs.tavus.io/sections/conversational-video-interface/pal/conversational-flow.md): Learn how to configure the Conversational Flow layer to fine-tune turn-taking and interruption handling behavior.
- [Speech-to-Text (STT)](https://docs.tavus.io/sections/conversational-video-interface/pal/stt.md): Configure the STT layer to select an STT model, improve transcription accuracy, and optimize for your target languages.
- [Large Language Model (LLM)](https://docs.tavus.io/sections/conversational-video-interface/pal/llm.md): Learn how to use Tavus-optimized LLMs or integrate your own custom LLM.
- [Perception](https://docs.tavus.io/sections/conversational-video-interface/pal/perception.md): Configure Raven for live camera, screen, and audio understanding, and request a separate post-call perception report.
- [Screen Share](https://docs.tavus.io/sections/conversational-video-interface/pal/screen-share.md): Let participants share their screen so your PAL can see and reason about what is on it in real time.
- [Overview](https://docs.tavus.io/sections/conversational-video-interface/skills/overview.md): Attach pre-built capabilities like internet search, presentations, and interactive UI to your PALs.
- [Presentation](https://docs.tavus.io/sections/conversational-video-interface/skills/presentation.md): Let your PAL walk participants through a slide deck from your Knowledge Base.
- [Magic Canvas Overview](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/overview.md): Let your PAL show interactive UI (questions, calendars, charts, and more) inside a live video conversation.
- [Magic Canvas Quickstart](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/quickstart.md): Attach the Magic Canvas skill to a PAL, render cards in the call, and receive every interaction on your webhook.
- [Canvas Components](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/components.md): Every card the PAL can show, and how to enable and configure them.
- [Component: question](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/components/question.md): Show a multiple-choice question card on screen and get a structured answer back.
- [Component: input](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/components/input.md): Collect a single typed value (a name, an email, a phone number) and receive it in your webhook.
- [Component: calendar](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/components/calendar.md): Let the user pick a date, a time slot, or a date range without leaving the conversation.
- [Component: scheduling_embed](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/components/scheduling-embed.md): Embed your Calendly booking page in the conversation and get a webhook when the meeting is booked.
- [Component: text](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/components/text.md): A read-only card of text (an optional heading and a body) shown alongside the live video.
- [Component: image](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/components/image.md): Show an image from your Knowledge Base or an allowlisted website on the canvas during a conversation.
- [Component: chart](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/components/chart.md): Render a bar, line, or pie chart on the canvas during a live conversation.
- [Component: alert](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/components/alert.md): Show a dismissible notice (info, success, warning, or error) inside a live conversation.
- [PAL Canvas Configuration](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/api/configuration.md): Full reference for the magic_canvas skill on PALs: attaching it, the component overlay, the skills API, and every validation error.
- [Canvas Interactions](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/api/interactions.md): Record card taps, receive canvas.interaction webhooks, and fetch Canvas interaction history.
- [Integration: Tavus-Hosted Embed and Widget](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/integrations/hosted.md): Render a Magic Canvas conversation from a single HTML tag, with no SDK code or rendering work.
- [Integration: @tavus/cvi-ui MagicCanvas](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/integrations/cvi-ui-sdk.md): Render Magic Canvas cards inside your own React app with the @tavus/cvi-ui MagicCanvas component: install, mount, and handle every interaction.
- [Internet Search](https://docs.tavus.io/sections/conversational-video-interface/skills/internet-search.md): Let your PAL answer questions with real-time web search.
- [Browser Use](https://docs.tavus.io/sections/conversational-video-interface/skills/browser-use.md): Let your PAL drive a live browser and narrate a guided flow while participants watch over screen share.
- [Overview](https://docs.tavus.io/sections/conversational-video-interface/pal/tools.md): How tool calling works on Tavus - LLM, perception, post-call, and system tools like end_call, plus delivery channels.
- [Tool Calling for LLM](https://docs.tavus.io/sections/conversational-video-interface/pal/llm-tool.md): Define reusable tools, attach them to any PAL, and deliver tool calls via app message or API call.
- [Tool Calling for Perception](https://docs.tavus.io/sections/conversational-video-interface/pal/perception-tool.md): Define vision, screen, and audio tools that fire when Raven sees or hears something during a conversation.
- [Post-Call Actions](https://docs.tavus.io/sections/conversational-video-interface/pal/post-call-tool.md): Run a tool automatically after a conversation ends - Tavus fills its open arguments from the transcript and perception analysis and delivers the result for you.
- [Tool Delivery](https://docs.tavus.io/sections/conversational-video-interface/pal/llm-tool-delivery.md): How tool calls reach your application - app messages to your frontend or HTTPS API calls to your backend.
- [Tool Authentication](https://docs.tavus.io/sections/conversational-video-interface/pal/llm-tool-auth.md): How Tavus authenticates to your endpoint when a tool is delivered via an API call.
- [MCP Connectors](https://docs.tavus.io/sections/conversational-video-interface/pal/mcp-connectors.md): Let a PAL delegate multi-step work to your MCP servers during a conversation, and follow it from your client.
- [Overview](https://docs.tavus.io/sections/conversational-video-interface/conversation/overview.md): Learn how to customize identity and advanced settings for a conversation to suit your needs.
- [Conversation Recordings](https://docs.tavus.io/sections/conversational-video-interface/quickstart/conversation-recordings.md): Store conversation recordings in your own S3, GCS, or Azure Blob storage. Federated identity - no secrets shared with Tavus.
- [Audio-Only Conversation](https://docs.tavus.io/sections/conversational-video-interface/conversation/customizations/audio-only.md): Start a conversation in audio-only mode, perfect for voice-only or low-bandwidth environments.
- [Call Duration and Timeout](https://docs.tavus.io/sections/conversational-video-interface/conversation/customizations/call-duration-and-timeout.md): Configure call duration and timeout behavior to manage how and when a conversation ends.
- [Closed Captions](https://docs.tavus.io/sections/conversational-video-interface/conversation/customizations/closed-captions.md): Enable closed captions for accessibility or live transcription during conversations.
- [Background Customizations](https://docs.tavus.io/sections/conversational-video-interface/conversation/customizations/background-customizations.md): Apply a green screen or custom background for a personalized visual experience.
- [Private Rooms](https://docs.tavus.io/sections/conversational-video-interface/conversation/customizations/private-rooms.md): Create authenticated conversations with meeting tokens for enhanced security.
- [Participant Limits](https://docs.tavus.io/sections/conversational-video-interface/conversation/customizations/participant-limits.md): Control the maximum number of participants allowed in a conversation.
- [Interaction Events](https://docs.tavus.io/sections/conver
Measured from the saved file
Linked document
UTF-8 text size
45.3 KB46432 bytes
Token count
10594o200k_base encoding
Markdown links
256254 distinct URLs
Fenced code blocks
0246 lines in the file
Size measures the captured UTF-8 text, not compressed transfer size. Tokens are counted with a named tokenizer; other models can count differently. Neither measure establishes context compatibility or answer quality.
Project H1
Tavus
Context architecture
How the content is divided
1 H2 sections organize 256 parsed Markdown links. 237 links point to .md or .markdown paths; 5 link occurrences use another hostname. Linked destinations have not been recursively validated.
H2 section
Links
UTF-8 bytes
Tokens
OpenAPI SpecsLine 241
2
141
33
Link counts exclude parsed fenced and inline code and resolve relative destinations against the final endpoint. Section sizes cover the text below each H2 until the next H2.
Embedded material
Code, installation and API examples
Fenced blocks
0 closed
Declared languages
No fenced blocks detected
Package-install commands
0 detected in fenced or inline code
Authentication examples
0 fenced lines match API-key, Authorization, Bearer, OAuth or access-token patterns
HTTP request examples
0 fenced lines match request-method patterns
Authentication links
1 links have authentication-related labels
Embedded example text and links to further documentation are counted separately. A code block does not establish that a sample runs or that an authentication flow succeeds.
Publisher prioritization
The Optional section
No H2 section named “Optional” was detected in this file.
The community proposal uses Optional by convention for secondary links that can be skipped for a shorter context. This label does not establish that everything outside it is critical, or that the material inside it is deprecated.
Machine-readable routes
MCP references and configuration text
MCP-related text was detected in the captured file. 3 links have MCP-related labels or paths. 0 fenced lines match mcpServers, mcp_servers or an mcp add command.
A verified text response establishes a companion at the checked URL. It does not prove the bundle contains all product documentation, that generation is automated, or that it fits an assistant’s context window.
Change evidence
What can be said about freshness
This capture establishes a baseline. Update frequency cannot be determined from one observation.
Linked release and migration material
1 links have release, changelog, deprecation or migration terms in their labels or paths. These provide places to investigate product changes separately.
File changes can reveal documentation edits or navigation changes. They do not by themselves measure product release velocity, accuracy, agent adoption or AI visibility.
Counts use a deterministic parser, not a spec-compliance grade. Unusual Markdown and indented code can fall outside its recognized patterns. See the llms.txt proposal and tokenizer implementation.