> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tavus.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Changelog

**October 30**

## Enhancements

* **Faster, more accurate Guardrails:** Guardrail checks are now 5x faster, while being more accurate with fewer false alarms. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/guardrails.md)
* **Faster, more accurate Objectives:** Objective checks are now 6x faster, while being more accurate in both status and transition steps. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/objectives.md)
* **Phoenix-4.5 Upgrades to Stock Faces:** We've upgraded an additional eight existing stock faces to Phoenix-4.5, bringing improved facial expressiveness and more natural movement. The upgraded replicas include Celine, Zane, Raj, Steve, Samantha, Nathan, Mary, and Victor.

**October 29**

## New Features

* **Emotion control in Pipecat:** With expressive conversation enabled (`PipelineParams(enable_expressive_conversation=True)`), `TavusTransport` and `TavusVideoService` set the PAL's facial emotion from the delivery the bot speaks in, such as `warm` to `happiness`. Requires `pipecat-ai` [1.15.0](https://github.com/pipecat-ai/pipecat/releases/tag/v1.15.0) or later. [Learn more](https://docs.tavus.io/sections/integrations/pipecat.md#advanced-avatar-emotions)

**October 27**

## New Features

* **Multilingual Knowledge Base:** Knowledge Base now supports documents and conversations in languages beyond English, using the same upload flow with no extra setup. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/knowledge-base.md)

## Enhancements

* **Improved Knowledge Base recall:** An upgraded retrieval model finds relevant passages more reliably, with a measured improvement of about 28 percentage points over the previous model on questions whose answers are in the documents. Existing documents have already been reprocessed, so no re-upload or configuration changes are required. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/knowledge-base.md)

**October 24**

## New Features

* **Emotion control in LiveKit:** Call `set_emotion` on the avatar session to set the emotion the PAL shows while it listens and speaks, or expose it as a tool so the LLM picks the emotion. Requires `livekit-plugins-tavus` [1.8.9](https://github.com/livekit/agents/releases/tag/livekit-agents%401.8.9) or later (Python) or `@livekit/agents-plugin-tavus` 1.9.4 or later (Node.js). [Learn more](https://docs.tavus.io/sections/integrations/livekit.md#advanced-avatar-emotions)

**October 23**

## New Features

* **Raven-2:** Our new perception model processes camera and screen content about 3x faster and keeps a visual history of the conversation, so PALs can refer back to objects, pages, and errors that are no longer visible. Set `layers.perception.perception_model` to `raven-2` to use it in your next conversation. Raven-1 remains the default. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/perception.md)

**October 15**

## New Features

* **Official Tavus ChatGPT plugin:** Build PALs directly in ChatGPT. The plugin runs on the Tavus MCP server, which uses the intelligence of Charlie from our PAL Maker behind the scenes, so system prompts follow the same prompt format and judgment as PALs built in platform. [Install it in ChatGPT](https://chatgpt.com/plugins/plugin_asdk_app_6aa1eba5a51881919a54483e9887bb66) or search for Tavus in the ChatGPT plugin store. [Learn more](https://docs.tavus.io/sections/agent-tools/mcp-server.md#use-tavus-in-chatgpt)
* **Face Studio:** Create high-quality, highly customized Phoenix-4.5 faces in minutes, right inside PAL Maker. Start from a video, a still photo, or a text prompt describing the face you want, or pick a starting point from the prompted face library. Then remix and adjust until the face looks right. [Try Face Studio](https://maker.tavus.io/dev/faces/create)
* **GPT Live 1.5 support:** Put a Tavus face on OpenAI's GPT Live 1.5 speech-to-speech model through Pipecat or LiveKit. GPT Live handles listening and speaking in one step, and Tavus renders a synchronized face from its audio. [Learn more](https://docs.tavus.io/sections/integrations/speech-to-speech.md)

## Enhancements

* **HEVC (H.265) support for face training:** Training videos for Phoenix-3, Phoenix-4, and Phoenix-4.5 faces can now use the HEVC (H.265) video codec in addition to H.264, so you can upload footage straight from most modern phones and cameras without re-encoding. [Video requirements](https://docs.tavus.io/sections/faces/phoenix-45-video-requirements.md)
* **Faster Audio-Only Mode:** Decreased utterance-to-utterance times for snappier audio-only conversations. Set [`audio_only`](https://docs.tavus.io/api-reference/conversations/create-conversation#body-audio-only) to `true` when you create a conversation. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/conversation/customizations/audio-only.md)

**October 13**

## New Features

* **Transfer assets between team accounts:** Users within a team can now transfer assets between accounts.

## Enhancements

* **Sparrow-2 is the default turn detection model:** PALs that omit `turn_detection_model` now use `sparrow-2`. `sparrow-1` is deprecated. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/conversational-flow.md)

**October 9**

## Enhancements

* **Improved PAL Maker:** Charlie now has access to the internet. He can pull from websites, look up recent information, and use the web to improve the PALs he creates for you.

**October 8**

## New Features

* **Official Agora support:** Tavus now officially supports Agora, a low-latency, real-time conversational WebRTC framework. [Learn more](https://docs.tavus.io/sections/integrations/agora.md)

**October 7**

## Enhancements

* **Phoenix-4.5 upgrades:** Better emotions for video-based faces and increased realism, with a reduction in micro twitches.

**October 5**

## Enhancements

* **Pipecat and LiveKit improvements:** Decreased audio-to-frame latency, plus other pipeline optimizations that improve the conversation experience. [Pipecat](https://docs.tavus.io/sections/integrations/pipecat.md) · [LiveKit](https://docs.tavus.io/sections/integrations/livekit.md)

**September 24**

## Enhancements

* **Account-scoped S3 recording keys:** Set `key_template` on an S3 `recording_storage` config and Tavus writes each recording under `tavus/<your_workspace_id>/…`, so you can restrict the recording role's permissions policy to your own prefix. Existing S3 setups continue to work unchanged. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/quickstart/conversation-recordings.md#amazon-s3)

**September 18**

## New Features

* **Dynamic Greeting:** Set `dynamic_greeting` on a PAL so conversations generate an opener from the PAL's context when there is no written greeting. Override the same field per conversation. Greeting priority is a written `custom_greeting`, then conversation `dynamic_greeting: true` (generates even if the PAL has a written greeting), then the PAL's `greeting`, then the PAL's `dynamic_greeting`. Generated in the first tag of the conversation `languages` override, otherwise the PAL's `languages`. [Create PAL](https://docs.tavus.io/api-reference/pals/create-pal.md) · [Create Conversation](https://docs.tavus.io/api-reference/conversations/create-conversation.md)

* **MCP connectors:** Give a PAL access to apps like Gmail, Google Calendar, and Linear, or plug in any third-party MCP server, so it can take real actions during a conversation. Register a connector, attach it to a PAL, and the PAL can send an email, schedule a meeting, or file a ticket on the user's behalf. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/mcp-connectors.md)

**September 17**

## New Features

* **Deployments page:** A new deployments page gives you an at-a-glance overview of every deployment in your account, including usage data for each one.

**September 16**

## New Features

* **Memory Stores:** Give each PAL a persistent memory for every participant. Tavus automatically learns an evolving profile and timeline, while optional pinned memories let you provide specific facts. Inspect and manage memory in PAL Maker or through the new Memory Stores API. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/memories.md)

## Enhancements

* **Speaker names in the transcript:** user turns in the [`application.transcription_ready`](https://docs.tavus.io/sections/webhooks-and-callbacks#application-callbacks) transcript now include `participant_name` when the conversation knows who spoke.

**September 15**

## Enhancements

* **Override objectives per conversation:** Pass `objectives_id` on [Create Conversation](https://docs.tavus.io/api-reference/conversations/create-conversation.md) to use a different set of objectives for a single call, without modifying the PAL. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/objectives.md#overriding-objectives-per-conversation)

**September 12**

## New Features

* **Microsoft Teams support:** PALs can now join Microsoft Teams calls in addition to Google Meet and Zoom. Invite the PAL's `@tavusinvite.com` address to a calendar event with a Teams link, or pass a Teams URL when creating a conversation. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/meetings.md)

**September 10**

## New Features

* **50 new Phoenix-4.5 stock faces:** 50 additional first-party faces built on Phoenix-4.5, including 28 Pro faces.

**September 9**

## New Features

* **16 new stock voices:** Added 16 new first-party voices to the Tavus catalog. Each voice is automatically localized to 16 languages and speaks all 42 supported languages. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/voices.md)

**September 2**

## New Features

* **Phoenix-4.5:** Newest face model.

  * Start using a face **in minutes** from a photo or a video, then it tunes in the background into an unwatermarked version.
  * Movement below the neck is more natural.
  * Glasses, jewelry, and hair in front of the shoulders are now allowed.
  * Extends support to animated human styles (cartoon, anime, Pixar-style).
  * The default when training a new face as of September 09, 2026.

  [Preview and status](https://docs.tavus.io/sections/faces/phoenix-45-preview-and-status.md) · [Choose a model](https://docs.tavus.io/sections/faces/which-training-path.md)

**September 1**

## Enhancements

* **Participant attribution on utterance events:** [`conversation.utterance`](https://docs.tavus.io/sections/event-schemas/conversation-utterance) and [`conversation.utterance.streaming`](https://docs.tavus.io/sections/event-schemas/conversation-utterance-streaming) now include `participant_id` and `participant_name`, using the same id [`conversation.joined`](https://docs.tavus.io/sections/event-schemas/conversation-joined-left) announced that participant under.

**August 28**

## New Features

* **SSO support:** Enterprise customers can now sign in to Tavus with single sign-on (SSO).
* **Teams:** Add team members to your account. Added members operate under your account's billing plan, while assets stay permission-bounded to the individual account that created them. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/faq.md#teams)

## Enhancements

* **Improved PAL publishing flow:** API edits through [Patch PAL](https://docs.tavus.io/api-reference/pals/patch-pal.md) now push directly to the live PAL, with better handling between Maker draft versions and the live version - including a [force push](https://docs.tavus.io/api-reference/pals/patch-pal.md#response-discarded-maker-changes) mechanism to discard a pending Maker draft. This makes the API simpler and PAL Maker easier to understand. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/editing-lifecycle.md)

## Changes

* **Publish PAL removed:** With API edits applying directly to the live PAL, the [Publish PAL](https://docs.tavus.io/api-reference/pals/publish-pal.md) endpoint has been removed from the API. PALs no longer have a separate published version.

**August 27**

## New Features

* **Joined/left interaction events:** New `conversation.joined` and `conversation.left` interaction events fire when a user or PAL enters or leaves a conversation, so your app can track who is present in real time. [Learn more](https://docs.tavus.io/sections/event-schemas/conversation-joined-left.md)

**August 26**

## New Features

* **PAL version history:** PAL Maker now keeps a version history, so you can see exactly what Charlie changed after he finishes making edits to your PAL.

**August 25**

## Enhancements

* **Improved PAL copying:** Copying and duplicating a PAL in PAL Maker is now a smoother process.

**August 24**

## Enhancements

* **Detailed conversation view:** The conversations page in PAL Maker now includes a detailed view for each conversation.

**August 21**

## New Features

* **100 new stock Tavus voices:** A new library of 100 first-party voices. Each voice is automatically localized to 16 languages and speaks all 42 supported languages. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/voices.md)
* **Custom voices:** Create and assign your own voices. Every voice is automatically localized to 16 languages and speaks all 42 supported languages. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/voices.md)
* **Languages on the PAL:** Assign the set of languages a PAL supports directly on the PAL. The same languages carry through every deployment and conversation the PAL is used in. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/language-support.md)
* **`tavus-auto` TTS engine:** A new default TTS engine that automatically routes each voice and language combination to the best-fit provider and model, so you no longer need to pick one by hand. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/tts.md)
* **Sparrow-2 available:** Opt in to the latest turn detection model today by setting `turn_detection_model` to `sparrow-2`. It becomes the default on September 8, 2026. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/conversational-flow.md)

**August 20**

## New Features

* **Updated Deployments landing pages:** Landing pages have more customization options and now support full-screen mode.

**August 19**

## Enhancements

* **Updated Deployments UI:** The Deployments interface has been redesigned to surface more information and make configuration clearer at a glance.

**August 18**

## Enhancements

* **PAL Maker revision history:** Revision history now shows exactly what changed and when, and lets you revert a PAL to any previous version or back to the current live version.

**August 17**

## Fixes

* **Magic Canvas reliability:** Fixed a set of reliability issues in Magic Canvas that caused components to occasionally fail to render or update mid-conversation.

**August 7**

## New Features

* **Magic Canvas image support:** The `image` component now renders images from your Knowledge Base or an allowlisted website during a conversation. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/components/image.md)
* **Auto-start recordings:** Set `auto_start_recording` on create-conversation and Tavus starts the recording about a second after the PAL joins, with no client-side call to `startRecording()`. Requires `recording_storage` and a Tavus-hosted room. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/quickstart/conversation-recordings.md#start-recording-automatically)

## Changes

* **New default LLM - `tavus-gemma-4`:** A faster, smarter, tuned version of `tavus-gemma-4` is now the recommended default. It makes for the snappiest conversations while staying strong on intelligence and tool use. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/llm.md#1-model)
* **Deprecated LLM:** `tavus-gemma-4-thinking` is deprecated in favor of the new tuned `tavus-gemma-4`.

**July 31**

## New Features

* **EU AI Act controls:** New AI disclosure controls on the PAL - `disclosure_type`, `verbal_disclosure`, and `visual_disclosure` - plus a per-conversation `policy: "eu"` that auto-applies the disclosure and switches Raven-1 to `limited` emotion recognition when PAL fields are left on `auto`. [Learn more](https://docs.tavus.io/sections/onboarding-guide/eu-ai-act.md)
* **Zoom support:** PALs can now join Zoom calls in addition to Google Meet. Invite the PAL's `@tavusinvite.com` address to a calendar event with a Zoom link and it joins on its own. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/meetings.md)

**July 21**

## New Features

* **Expanded Tavus-Hosted LLM Selection:** Added support for new Tavus-hosted LLMs: `tavus-gemma-4`, `tavus-gpt-5.6-sol`, and `tavus-gpt-5.6-terra`. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/persona/llm.md#tavus-hosted-models)

## Changes

* **Deprecated LLMs:** The following Tavus-hosted LLMs are now deprecated: `tavus-glm-4.7`, `tavus-gpt-oss`, `tavus-claude-haiku-4.5`, `tavus-gpt-5.2`, and `tavus-gemini-3-flash`. PALs still set to `tavus-glm-4.7` route to `tavus-gpt-5.6-terra`.

**June 26**

## New Features

* **Zero Data Retention:** Account-level and programmatic support for Zero Data Retention is now available, enabling customers to opt out of data persistence across Tavus services.

**June 24**

## Enhancements

* **Performance:** Improved LiveKit pipeline performance by 15%!

**June 19**

## New Features

* **PAL Maker:** Describe a PAL in plain language and have it configured end to end (system prompt, face, objectives, guardrails, tools, knowledge base, and advanced settings), then talk to it in the same session. Removes the need to configure each layer by hand. [Learn more](https://maker.tavus.io/dev)
* **Starter Templates:** Fork a complete, live experience into your account with no code (Interviewer, SDR, Medical Intake, Trivia Host), or open the GitHub repo behind it. [Learn more](https://maker.tavus.io/dev)
* **Meetings (Google Meet):** Send a PAL into a Google Meet call. It joins on its own past the admit prompt, sees and hears the room, and can run the meeting. Invite it via calendar invite or pass a meeting URL when creating a conversation. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/meetings.md)
* **Magic Canvas:** A shared interactive surface rendered on top of the video stream. The PAL surfaces the right component at the right moment. Six components at launch: question, text, input, calendar, chart, and alert. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/magic-canvas/overview.md)
* **Presentation Mode:** Upload a deck or source material, Tavus indexes it, and the PAL presents from it, takes questions, and pulls the right slide on demand. Supports two modes: Walk the deck (end to end) and On-demand. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/skills/presentation.md)
* **Embeddable Widgets:** Drop a PAL onto any site with a single line of code, scoped and auto-provisioned to your account, with no SDK. Place it in any corner or inline, and customize branding to match your site. [Learn more](https://docs.tavus.io/sections/deployments/widget.md)
* **Landing Pages:** Deploy a PAL on a ready-made hosted landing page, no setup required. [Learn more](https://docs.tavus.io/sections/deployments/landing-page.md)
* **Internet Search:** Give a PAL real-time web search during a conversation. Attach the `internet_search` skill - pure on/off, no configuration. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/skills/internet-search.md)

## Enhancements

* **Tool Calling:** Tools can now call any third-party API (including GraphQL) with custom auth and request bodies, no middleware. New on-resolve modes (fire and forget, pull into context, generate a response) control how results enter the conversation, and tools are reusable across PALs. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/pal/tools.md)

**June 12**

## Enhancements

* **Meet Rivian, Tiffany, and Brian:** Our 3 new stock replicas!
* **Improved Image to Replica uploading flow:** Get more detailed and accurate errors before training submission.

**June 5**

## New Features

* **Guardrails as first-class primitives:** Guardrails are now standalone resources that can be created, edited, and reused across multiple PALs. Compose individual guardrails into sets via `guardrail_tags`. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/guardrails.md)
* **Choose your delivery channel:** Guardrail violations can now be delivered via [Interaction Events](https://docs.tavus.io/sections/conversational-video-interface/interactions-protocols/overview.md) (in-call `app-message`) or [Webhooks](https://docs.tavus.io/sections/webhooks-and-callbacks.md#guardrail-callbacks) (server-side `callback_url`) - or both.
* **`guardrail_uuid` on violations:** Triggered events and callbacks now include `guardrail_uuid`, so you can identify exactly which individual guardrail was violated.

**June 3**

## New Features

* **Azure TTS support:** Azure is now available as a TTS engine, expanding language coverage for multilingual and localized voice output. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/persona/tts.md)

**May 22**

## New Features

* **Individually addressable guardrails:** Guardrails are now first-class resources. Create, attach, edit, and delete each guardrail independently via the [Guardrails API](https://docs.tavus.io/api-reference/guardrails/create-guardrails.md). Bundle guardrails by tag and reference them on a PAL via `guardrail_ids` or `guardrail_tags`. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/guardrails.md)
* **Node.js plugin support for LiveKit integration:** The LiveKit Agents integration now supports Node.js via the `@livekit/agents-plugin-tavus` plugin, in addition to Python. Install with `npm install @livekit/agents @livekit/agents-plugin-tavus`. [Learn more](https://docs.tavus.io/sections/integrations/livekit.md#node-js)
* **Real-time event timestamps:** Every interaction event (`conversation.utterance`, `conversation.started_speaking`, etc.) now carries a `timestamp` field delivered in real time on the respective Interaction event. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/interactions-protocols/overview.md)
* **Transcript utterance timestamps:** End-of-call transcripts now carry per-turn timing. The [`application.transcription_ready`](https://docs.tavus.io/sections/webhooks-and-callbacks#conversation-callbacks) webhook (and the same payload nested under `events` in the verbose [`GET /conversations/{id}?verbose=true`](https://docs.tavus.io/api-reference/conversations/get-conversation) response) now includes `timestamp` (Unix epoch float, seconds - same field name as live interaction events), `seconds_from_start`, `duration` (seconds, float - same field name as `conversation.stopped_speaking.duration`), and `inference_id` (on assistant turns) on each transcript entry.

## Enhancements

* **New, and more detailed, error messages when a face fails:** [Learn more](https://docs.tavus.io/sections/errors-and-status-details.md#face-training-errors)

**May 20**

## New Features

* **AI Image Fixer API support:** [Create Face](https://docs.tavus.io/api-reference/phoenix-replica-model/create-replica.md) now accepts an `auto_fix_training_image` property. Set it to `true` to use Tavus's AI Image Fixer to instantly fix any uploaded image to fit our requirements, eliminating the need for editing or recapturing photos. [Learn more](https://docs.tavus.io/sections/replica/train-with-an-image.md#ai-image-fixer)

**May 12**

## New Features

* **AI Image Fixer:** Instantly fix any uploaded image to fit our requirements, eliminating the need for editing or recapturing photos.

**May 11**

## New Features

* **Frame Checker (Video to Face):** Get instant feedback, before recording, on whether your camera setup meets our requirements.

## Enhancements

* **Less footage required (Video to Face):** We now only require 1 minute of video, down from the 2 minutes previously needed.
* **Simpler in-portal recording flow (Video to Face):** A streamlined recording experience in the portal to help you capture high-quality recordings.

**May 6**

## New Features

* **Image to Face:** Build a face from a single still. Drop in a photo, illustration, or brand mascot.

**May 8**

## New Features

* **Voice Activity Detector improvements:** Resulting in a smoother conversational experience in noisy environments. This is automatically rolled out to all users.
* **Expanded Recording Storage Support:** Conversation recordings can now be delivered to Google Cloud Storage (GCP) and Azure Blob Storage, in addition to AWS S3. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/quickstart/conversation-recordings.md#set-up-your-storage)

## Enhancements

* **Tavus Components Library Updates:** Improved audio-video sync, plus new chat components and closed captions with streaming support. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/component-library/components.md)

**May 1**

## New Features

* **Wake Phrase:** PALs can now stay silent until they hear a specific phrase, similar to how voice assistants like Siri or Alexa work. Configure it via the `wake_phrase` parameter in the Conversational Flow layer. The PAL still hears everything that is said and responds with full conversation history once the wake phrase is detected. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/persona/conversational-flow.md#5-wake_phrase)

**April 30**

## New Features

* **Idle Engagement:** Faces can now proactively re-engage by speaking to the user after a period where the user is silent. Eagerness of this feature can be configured via the `idle_engagement` parameter in the Conversational Flow layer. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/persona/conversational-flow.md#6-idle_engagement)

**April 29**

## New Features

* **Speaking Events:** Two new events - `conversation.started_speaking` and `conversation.stopped_speaking` - fire for both the PAL and the user with a `role` field (`"pal"` or `"user"`) identifying the speaker. Tavus also sends legacy duplicate events with `"replica"` instead of `"pal"`. `conversation.stopped_speaking` includes an `interrupted` boolean and a `duration` field (in seconds). [Learn more](https://docs.tavus.io/sections/event-schemas/conversation-started-stopped-speaking.md)
* **Conversation Diagnostics:** A new diagnostics surface for inspecting what happened in a conversation - including packet loss, network connection, FPS, and more - designed to make debugging significantly faster. Click on any conversation in the PAL Maker to access its diagnostics page.

**April 17**

## Enhancements

* **Non-Interruptible Custom Greetings:** Custom greetings now finish entirely before users can interrupt speech. Previously, participants could talk over a `custom_greeting`; now the PAL completes the greeting before it begins listening. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/faq.md#general-tavus-q-and-a:can-participants-interrupt-the-faces-greeting)
* **Improved Turn-Taking Latency:** Significant TTS optimizations reduce turn-taking latency, resulting in faster and more natural back-and-forth during conversations.

**April 10**

## New Features

* **Streaming Utterance Event:** A new `conversation.utterance.streaming` event progressively reports what has been said during a conversation turn for both face and user utterances. Use it to power closed captioning and build accurate transcripts - especially when a user interrupts the face, since the streaming event reflects only the words actually spoken rather than the full LLM response. [Learn more](https://docs.tavus.io/sections/event-schemas/conversation-utterance-streaming.md)
* **Pronunciation Dictionaries:** Define custom pronunciation rules so your PAL says brand names, technical terms, acronyms, and foreign words exactly right. Choose between simple alias substitution (e.g., "Tavus" → "TAH-vus") or precise IPA phonetic notation. Create a dictionary once and attach it to a PAL via the TTS layer - any updates automatically propagate to all linked PALs with zero extra latency at conversation time. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/persona/pronunciation-dictionaries.md)

**April 7**

## New Features

* **Voice Isolation:** Filter background noise from participant audio to improve conversation quality. Configure it via the `voice_isolation` parameter in the Conversational Flow layer. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/persona/conversational-flow.md#4-voice_isolation)

**April 3**

## Changes

* **Chat Interrupt History:** PALs now know when they have been interrupted. This allows the PAL to pick back up where it left off, and also improves objectives adherence.

**March 27**

## New Features

* **Expanded ASR Model Selection:** You can now choose from five specialized speech-to-text engines via the `stt_engine` parameter. New models include `tavus-parakeet`, `tavus-soniox`, `tavus-whisper`, and `tavus-deepgram-medical`. Use `tavus-auto` to automatically route to the best model for each conversation. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/persona/stt.md)

**March 19**

## Enhancements

* **30% Faster Phoenix-4 Boot Time:** Phoenix-4 conversations now boot 30% faster, significantly reducing the time from conversation creation to readiness.

## Changes

* **`conversation.replica_interrupted` Event Removed:** The `conversation.replica_interrupted` application message has been removed from interaction events. This event was deprecated in a previous backend update. Use `conversation.replica.stopped_speaking` with the `interrupted: true` property to detect interruptions instead.
* **`duration` and `interrupted` Fields on Face Stopped Speaking:** The `conversation.replica.stopped_speaking` event now includes a `duration` field (how long the PAL spoke in seconds) and an `interrupted` field (`true`/`false`) indicating whether the PAL was interrupted by the user. [Learn more](https://docs.tavus.io/sections/event-schemas/conversation-replica-started-stopped-speaking)

**March 17**

## New Features

* **Event Ordering and Turn Tracking:** All server-broadcasted interaction events now include `seq` and `turn_idx` fields. `seq` is a globally monotonic sequence number for ordering events that may arrive out of order, and `turn_idx` groups related events from the same conversation turn. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/interactions-protocols/overview.md#event-ordering-and-turn-tracking)

**March 11**

## Enhancements

* **30% Faster Phoenix-4 Boot Time:** Phoenix-4 conversations now boot 30% faster, significantly reducing the time from conversation creation to readiness.

**March 5**

## Enhancements

* **EU ElevenLabs BYOK Support:** Customers can now bring their own ElevenLabs API key from EU-region accounts.

**February 19**

## Enhancements

* **Improved Knowledge Base Retrieval:** Optimized underlying infrastructure to improve utterance to utterance response times, particularly when `rag_search_quality` is set to `quality`.

**February 17**

## New Features

* **Expanded Tavus-Hosted LLM Selection:** Added new Tavus-hosted LLM options including models from Gemini, Claude, and GPT families. `tavus-gpt-oss` is recommended as the default. Legacy models `tavus-gpt-4.1`, `tavus-gpt-4o`, and `tavus-gpt-4o-mini` are now deprecated. [Learn more →](https://docs.tavus.io/sections/conversational-video-interface/persona/llm.md#tavus-hosted-models)
* **Visual RAG:** CVI now supports visual retrieval-augmented generation. Upload custom image explanations that are matched and queried via vision embeddings, giving your PAL richer visual context during conversations.

**February 12**

## Changes

* **PAL**`context`**Field Deprecated:**  The `context` field has been deprecated in favor of a unified `system_prompt` field. Existing `context` values have been automatically merged into system prompts. The API remains backward compatible, but we recommend using **only** `system_prompt` going forward.

**February 11**

## New Features

* **Raven-1 Perception Model:** Introduced Raven-1, a multimodal perception model with audio emotion analysis and enhanced visual awareness. Raven-1 captures user emotion from audio in real time (sub-100ms audio perception latency), enabling PALs to respond with greater emotional intelligence. The model is now the default for all new PALs. Enable it by setting `perception_model_name` in your PAL configuration. [Learn more →](https://www.tavus.io/post/raven-1-bringing-emotional-intelligence-to-artificial-intelligence)
* **Private Rooms:** Require authentication to join conversations for enhanced security. When enabled, we return a JWT meeting token that users must include when entering the room. [Learn more](https://docs.tavus.io/sections/conversational-video-interface/conversation/customizations/private-rooms.md)

**February 9**

## Enhancements

* **Upgraded Transcription Engine:** Upgraded transcription engine with 3x improvements in word error rates (WER).

**January 17**

## New Features

* **Website Crawling for Knowledge Base:** You can now enable link crawling when creating knowledge base documents. Configure crawl `depth` and `max_pages` to automatically discover and ingest content from linked pages. Additionally, existing crawled documents can now be recrawled to keep knowledge base content up to date.

**January 8**

## Changes

* **PlayHT TTS Removed:** PlayHT has been fully removed as a supported TTS engine. All PALs previously using PlayHT should migrate to Cartesia or ElevenLabs.

**December 18**

## New Features

* **Hard Delete for Conversations:** Conversations can now be permanently deleted via the API using the `hard=true` query parameter. Use this for GDPR compliance or data cleanup workflows.

## Enhancements

* **Default TTS Model Updated to Sonic-3:** The default text-to-speech model has been updated to Sonic-3 across all new PALs, delivering improved voice quality and naturalness.
* **LiveKit Connection Stability:** Extensive reliability improvements to the LiveKit-based transport layer, including fixes for connection timeouts, track publishing hangs, event loop starvation, and ping timeout issues.

## Changes

* **Default LLM Migrated to `tavus-gpt-oss`:** The default LLM for all new PALs is now `tavus-gpt-oss`. All remaining `tavus-llama-4` PALs have been automatically migrated. Legacy Tavus-Llama model references have been removed.

**November 27**

## New Features

* **LLM Temperature & Top-P Parameters:** You can now configure `temperature` and `top_p` parameters for both Tavus-hosted LLMs and custom LLMs via the `extra_body` field in your PAL's LLM configuration. [Learn more →](https://docs.tavus.io/sections/conversational-video-interface/persona/llm.md#4-extra-body)

## Enhancements

* **Text Echo Language Accuracy:** Text echoes now correctly use the input language for conversion, improving accuracy in multilingual conversations.

**September 17**

## New Features

* **Test Mode for Conversations:** You can now start conversations in test mode, where the PAL does not join. Validate your setup, integrations, and conversational flows without incurring costs or using concurrency slots. Set `test_mode: true` when creating a conversation. [Learn more →](https://docs.tavus.io/api-reference/conversations/create-conversation.md#body-test-mode)

**September 7**

## Enhancements

* **Fuzzy Search for PALs:** Search now supports fuzzy matching for PALs, allowing users to find results based on partial matches of UUIDs or names.

**August 15**

## New Features

* **Memories:** CVI now remembers context across conversations. Every conversation builds on the last with full context and time/date awareness, enabling use cases like adaptive tutoring, mentorship, and recurring consultations. [Learn more →](https://docs.tavus.io/sections/conversational-video-interface/memories.md)
* **Knowledge Base (RAG):** Bring your own data to conversations instantly. Upload documents or links and get grounded answers with \~30ms retrieval latency. Power AI recruiters, support agents, travel guides, and more with domain-specific knowledge. [Learn more →](https://docs.tavus.io/sections/conversational-video-interface/knowledge-base.md)
* **Objectives & Guardrails:** Define clear goals, branching logic, and measurable outcomes for your PALs while keeping conversations safe, compliant, and on-brand. Ideal for complex workflows and regulated industries. [Learn more →](https://docs.tavus.io/sections/conversational-video-interface/persona/objectives.md)
* **PAL Builder:** A guided creation flow in the PAL Maker to shape AI PALs with goals, behaviors, and style - then test or launch within minutes.

**July 28**

## New Features

* **Events Console:** A new events console in the PAL Maker lets you monitor everything happening during a conversation in real time - from message flows to system activity.
* **Conversation Transcripts & Perception Analysis:** View full conversation details directly in the PAL Maker, including transcripts with speaker roles and perception analysis showing how your AI PAL sees, hears, and responds.

**July 25**

## New Features

* **PAL Layer Controls:** Enable or disable layers like Sparrow directly within a PAL and adjust sensitivity settings in real time from the PAL Maker side panel.
* **PAL editing in PAL Maker:** We've added new editing capabilities to help you refine your PALs more efficiently. You can now update system prompt, context, and layers directly in our PAL Maker, plus duplicate existing PALs to quickly create variations or use them as starting points for new projects. Find these new features in your PAL Library at maker.tavus.io/dev.

**July 24**

## Enhancements

* **Interaction Events Playground Improvements:** Major updates to the Interaction Events Playground including correct `properties.context` format and append vs overwrite toggle.

**July 23**

## New Features

* **Multilingual Settings in PAL Maker:** You can now specify the language of a conversation directly in the PAL Maker, including a new multilingual option for dynamic, real-world interactions.

**July 22**

## New Features

* **Llama 4 Support:** Your PAL just got even smarter, thanks to Meta's Llama 4 model 🧠 You can start using Llama 4 by specifying `tavus-llama-4` for the LLM `model` value when creating a new PAL or updating an existing one. Click [here](https://docs.tavus.io/sections/conversational-video-interface/persona/llm.md#tavus-hosted-models)

  to learn more!

**July 15**

## New Features

* **React Component Library:** Developers can build with Tavus even faster now with our pre-defined components 🚀 Click [here](https://docs.tavus.io/sections/conversational-video-interface/component-library/overview.md)

  to learn more!

**June 27**

## New Features

* **Multilingual Conversation Support:** CVI now supports dynamic multilingual conversations through automatic language detection. Set `properties.language` to "multilingual" and CVI will automatically detect the user's spoken language and respond in the same language using ASR technology.
* **Audio-Only Mode:** CVI now supports audio-only conversations with advanced perception (powered by Raven) and intelligent turn-taking (powered by Sparrow-1). Set `audio_only=true` in your create conversation request to enable streamlined voice-first interactions.

**June 20**

## Enhancements

* **Fixed CVI responsiveness issue:** Resolved an issue where CVI would occasionally ignore very brief user utterances. All user inputs, regardless of length, now receive consistent responses.
* **Expanded tavus-llama-4 context window:** Increased maximum context window to 32,000 tokens. For optimal performance and response times, we recommend staying under 25,000 tokens.

**June 3**

## Enhancements

* Reduced conversation boot time by 58% (p50).

**May 28**

## Changes

* Added a new recording requirement to [Training from a video](https://docs.tavus.io/sections/replica/train-with-a-video.md)

  : Start the talking segment with a big smile.

## Enhancements

* Added [echo](https://docs.tavus.io/sections/event-schemas/conversation-echo.md)

  and [respond](https://docs.tavus.io/sections/event-schemas/conversation-respond.md)

  events to conversational context.

**May 17**

## Enhancements

* **Major Phoenix 3 Enhancements for CVI**:
  * Increased frame rate from 27fps to 32fps, significantly boosting smoothness.
  * Reduced Phoenix step's warm boot time by 60% (from 5s to 2s).
  * Lipsync accuracy improved by \~22% based on AVSR metric.
  * Resolved blurriness and choppiness at conversation start.
  * Enhanced listening mode with more natural micro expressions (eyebrow movements, subtle gestures).
  * Greenscreen mode speed boosted by an additional \~1.5fps.
* **Enhanced CVI Audio Quality**: Audio clicks significantly attenuated, providing clearer conversational audio.
* **Phoenix 3 Visual Artifacts Fix**: Resolved visual artifacts in 4K videos on Apple devices, eliminating black spot artifacts in thumbnails.

**May 9**

## New Features

* Launched [LiveKit Integration](https://www.tavus.io/post/building-real-time-ai-video-agents-with-livekit-and-tavus)

  : With Tavus video agents now integrated into LiveKit, you can add humanlike video responses to your voice agents in seconds.
* [PAL API](https://docs.tavus.io/api-reference/pals/patch-pal.md)

  : Enabled patch updates to PALs.

## Enhancements

* Resolved TTS (Cartesia) stability issues and addressed hallucination.
* **Phoenix 3 Improvements**:
  * Fixed blinking/jumping issues and black spots in videos.
  * FPS optimization to resolve static and audio crackling.

**April**

## Enhancements

* **Face API**:
  * Enhanced Error Messaging for Training Videos.
  * Optimized Auto QA for Training Videos.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.