What is a Persona Agent?
A Persona Agent is a custom AI assistant with a personality, expertise, and communication style that you define. Unlike a generic chatbot, a Persona remembers your preferences and context across every conversation — and your configuration applies to every AI model in Wysor.
What you can customize
Role and expertise
Define what the agent specializes in. An executive assistant, a brand voice guardian, a technical advisor, a legal consultant — anything. The agent's behavior is shaped by the role you give it.
Communication style
Choose how the agent speaks: professional, friendly, formal, or consultative. This controls tone, formality, and response structure.
Master prompt
The master prompt is the heart of your Persona. Write detailed instructions: what it should always do, what it should never do, how it should handle uncertainty, what vocabulary to use or avoid, how to format responses. The master prompt is injected automatically into every conversation and can be updated at any time.
If you're not sure what to write, use the Improve with AI button — Wysor will suggest a refined version of your prompt based on the role you set.
Works with every AI model
Your Persona stays consistent regardless of which AI model powers the response. Start a conversation with Claude, switch to GPT-5 for a specific question, move to Gemini for research — the personality, context, and instructions stay the same throughout.
How to create one
- Go to Agents in your sidebar and click Create agent.
- Select Persona as the agent type.
- Give it a name and pick an avatar.
- Choose a role template or start from scratch.
- Write (or improve) the master prompt that defines the agent's behavior.
- Save, then start chatting. You can refine the configuration any time — changes apply immediately.
Supercharge Skills
Skills are plug-and-play behavior modules you can add to any Persona. Instead of writing prompt engineering from scratch, pick skills from the library and stack them:
- Structured Output — responses come back as clean tables, bullet lists, or custom formats
- Citation Mode — every claim backed by a source
- Tone Enforcement — stays in character even under adversarial prompts
- Multi-Format Export — output in Markdown, JSON, or plain text on demand
Skills layer on top of your master prompt. You can add or remove them at any time.
Memory
Personas can remember facts about you across conversations — preferences, context, project details. Toggle memory on from the agent settings. Saved memories are visible, editable, and deletable any time.
Privacy
Your master prompt and conversation history are encrypted with AES-256. No AI provider trains on your data. You can delete any Persona or individual conversation at any time.