> Index of Phersa pages for AI agents: https://phersa.com/llms.txt

# Personas

A persona is the AI person who stars in your videos. It is one reference photo, used in every clip so the same person appears throughout, plus an optional voice and a few notes about their world.

## Create a persona

1. Open **Personas** and choose **New Persona**.

2. Give the persona a name, then drop one photo of them on **Drop a portrait**, or click it to pick a file.

3. Optionally add a **Voice** and write **World** notes. You can change all of it later.

Upload the photo here in the app. Your [AI agent](https://phersa.com/docs/mcp) can set a persona's photo from a file already in Assets or Results, or from a public https link, but not from a file you drop into the chat.

## The photo is the identity: same face in every video

Phersa doesn't describe your persona in words. It sends the photo as the reference image for the first clip of every project, so the same face carries through your video.

Use a sharp, well-lit photo with the face clearly visible. Need one? The [persona ideas](https://phersa.com/personas) page has example personas with the full prompt that made each photo.

## Voice

Paste the Voice ID of a voice in your ElevenLabs library and choose **\+ Add**. Voice Swap in Create uses it to give each clip your persona's voice, so it is required for Voice Swap. You need an ElevenLabs key connected in Settings; see [BYOK](https://phersa.com/docs/byok). Your AI agent can also list your ElevenLabs voices over [MCP](https://phersa.com/docs/mcp).

## World

World notes describe the persona's setting and niche, up to 600 characters: where they live, what they talk about, what their videos look like. Your AI agent reads them over [MCP](https://phersa.com/docs/mcp) when it writes clips.

## Results

Next to the persona's identity, the **Results** column shows every finished take generated with them. Deleting a persona keeps those takes in the main Results tab.

Source: https://phersa.com/docs/personas