> For the complete documentation index, see [llms.txt](https://project-45.gitbook.io/tavus/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://project-45.gitbook.io/tavus/conversational-video-interface/creating-a-replica.md).

# Creating a Replica

#### [​](https://docs.tavus.io/sections/conversational-video-interface/creating-a-replica#the-replica-is-the-talking-head-or-the-face-to-the-digital-twin)The replica is the ‘talking head’ or the face to the digital twin. <a href="#the-replica-is-the-talking-head-or-the-face-to-the-digital-twin" id="the-replica-is-the-talking-head-or-the-face-to-the-digital-twin"></a>

The first step to using CVI is selecting a replica to use as the ‘face’ of your digital twin. Tavus has stock replicas you can use as well as the ability to create custom replicas via the API or the portal.

### [​](https://docs.tavus.io/sections/conversational-video-interface/creating-a-replica#stock-replicas)Stock Replicas <a href="#stock-replicas" id="stock-replicas"></a>

You can get started quickly be using one of our stock replicas. We have a few replicas that we recommend for conversational usage:

Reccomended Stock Replicas for CVI

### [​](https://docs.tavus.io/sections/conversational-video-interface/creating-a-replica#custom-replica)Custom Replica <a href="#custom-replica" id="custom-replica"></a>

You can use a custom or ‘personal’ replica to be the face of your digital twin. If you have already created a custom replica for video generation you can reuse that replica for CVI. However, what looks good for video generation does not necessary look good for conversational (CVI). We recommend following the instructions for [Creating a good replica for CVI](https://docs.tavus.io/sections/replicas/personal-replicas) for the best results.

### [​](https://docs.tavus.io/sections/conversational-video-interface/creating-a-replica#what-makes-for-a-good-replica-for-cvi)What makes for a good replica for CVI? <a href="#what-makes-for-a-good-replica-for-cvi" id="what-makes-for-a-good-replica-for-cvi"></a>

#### [​](https://docs.tavus.io/sections/conversational-video-interface/creating-a-replica#silent-frames)Silent frames <a href="#silent-frames" id="silent-frames"></a>

The main difference between using a replica for video generation vs CVI is that videos don’t have long periods of pauses, whereas during a conversation you take turns, therefore the replica sits in silent listening or waiting. This can look odd if you try to use a replica that is meant for video generation because the replica might move unnaturally during these periods of silence.

#### [​](https://docs.tavus.io/sections/conversational-video-interface/creating-a-replica#casual-low-production-environment)Casual/low-production environment <a href="#casual-low-production-environment" id="casual-low-production-environment"></a>

For most use cases CVI is supposed to feel like a 1:1 call. It should feel like you’re jumping on a Zoom call with someone. This means that the setting and environment should feel like a Zoom call, not a studio environment. A webcam at a desk for example will feel more natural than an awkward replica that is standing the entire time. Users don’t expect you to be in a studio every time you’re on a Zoom call and it can actually detract from the experience. This doesn’t mean you can’t shoot in studio, it just means that the studio setting itself should look casual as well.

Learn more about [Creating a good replica for CVI](https://docs.tavus.io/sections/replicas/personal-replicas) for the best results.