A human-like AI personal assistant from the US company Tavus, complete with voice, expression, and gesture. You can move seamlessly between text chat, voice calls, and video calls, and its defining trait is that it reads your facial expression and tone of voice through the camera as it responds. It remembers what you have talked about along with context from your schedule and email, and it reaches out on its own with reminders and check-ins. PALs launched in November 2025 and is available in the browser as well as through the iOS app “PALs by Tavus: AI Companions.”
Tavus originally built a real-time conversational video platform for developers (CVI: Conversational Video Interface), and PALs is that same technology repackaged for everyday users.
Key Features
- Perception of expression, tone, and surroundings: The Raven-1 perception model interprets facial expressions, gaze, tone of voice, and environmental context through the camera in real time. That makes it possible for the assistant to pick up on how you seem and speak up before you put it into words
- A face that behaves naturally on video calls: The Phoenix-4 rendering model generates context-appropriate expressions, emotions, and movement in real time. The result feels different from avatars that simply add lip-sync to a still image
- Human-like turn-taking: The Sparrow-1 conversation model judges not only what you are saying but how you are saying it and whether you are about to finish, then decides when to speak. That avoids the clipped interruptions and awkward pauses common to voice assistants
- Seamless switching between text, voice, and video: Text while you are out, voice when your hands are busy, video when you want to talk things through ─ all with the same companion and the same context
- Persistent memory and ongoing learning: It remembers past conversations, goals, and concerns, so the next conversation picks up where the last one left off. The design is meant to accumulate understanding of you the more you use it
- Calendar and email integration with proactive outreach: It connects to your calendar, email, and other apps to grasp context, and offers reminders and status checks without being asked. Higher tiers also include early access to MCP (Model Context Protocol) integration
Pricing
PALs (the consumer product) gives you unlimited messaging, with a set monthly allowance of voice and video call minutes per plan.
| Plan | Monthly price | What’s included |
|---|---|---|
| Free | $0 | Unlimited messaging, 15 minutes/month of voice and video calls, 30+ languages |
| Plus | $20 | Unlimited messaging, 150 minutes/month of voice and video calls, MCP early access |
| Max | $50 | Unlimited messaging, 500 minutes/month of voice and video calls, MCP early access |
Annual plans and any associated discounts are not listed on the public pricing page; they are shown in the in-app billing screen. Note that the developer-facing CVI (conversational video API) is on a separate pricing structure, with a monthly base fee that includes an allowance of conversation minutes plus pay-as-you-go billing beyond it.
Pricing is current as of August 2026. Check the official site for the latest information.
Pros & Cons
✅ Pros
- Because it factors in your expression and tone through the camera, it feels more attentive than text-first AI assistants
- You can move between text, voice, and video while keeping the same context, making it easy to match the format to the situation
- It remembers conversations and goals, so you do not have to re-explain yourself, and it fits better the longer you use it
- Calendar and email integration lets it surface reminders proactively
- Even the free plan offers unlimited messaging plus 15 minutes of calls per month to try it out
- The iOS app makes it easy to reach for on a phone day to day
⚠️ Cons
- Voice and video calls are metered, and the free plan’s 15 minutes per month is only enough to get a feel for it. Regular call use effectively requires a paid plan
- Getting value from expression reading means pointing a camera at yourself, which is not practical in every setting
- Calendar and email integration is convenient, but it widens the range of personal data you hand over, so review data handling before using it for work
- Annual plans and discounts are not published on the pricing page, which makes comparison before signing up harder
- Quality in Japanese ─ both naturalness and real-time responsiveness ─ may not match the English experience
- The product is still young, so features and pricing may continue to change
Comparison with Similar Services
| Criteria | PALs | Replika | Character.AI | ChatGPT (voice mode) |
|---|---|---|---|---|
| Provider | Tavus | Luka | Character Technologies | OpenAI |
| Main use | An everyday companion for talking things through and managing your schedule | Emotional companionship and casual chat | Roleplay and chat with characters | General-purpose assistant |
| Video calls | Yes (real-time generated expressions) | Limited (3D avatar based) | No | No (screen sharing and camera input available) |
| Expression and tone reading | Yes | Limited | No | Reflects tone of voice; visual expression reading is limited |
| Calendar and email integration | Yes | No | No | Available via connectors |
| Free plan | Yes (15 call minutes/month) | Yes | Yes | Yes |
For general-purpose work, an all-rounder like ChatGPT has the edge, and Character.AI offers far more character variety. What PALs brings is talking face to face on video and being understood through expression and tone as well as words.
Who Is It For
- People who want an AI they can talk to face to face rather than through text
- Anyone who wants a companion that remembers schedule and email context and follows up with reminders and check-ins
- Language or interview practice, or any use case where expression and delivery should be part of the exchange
- People who want to switch between voice on the move and video when they can sit down
- Anyone curious to experience Tavus’s conversational video technology as a finished product before touching the developer API
Summary
PALs takes the real-time conversational video technology Tavus has built up and turns it into a personal companion you can use directly. Reading expression and tone, talking face to face, and carrying memories forward into the next conversation add up to an experience qualitatively different from text-first AI assistants. That said, voice and video calls are metered, and the free plan’s 15 minutes is no more than a trial. A practical approach is to start free, judge how natural the calls feel, and move to Plus or above once you decide it fits into your day.