PERSONASTACK DOCUMENTATION

AI providers

Your personas use saved provider connections to reach their AI models. A connection can serve several personas in the same workspace, so review its assignments before changing it.

Start here

Start with an existing account or API key. Connect it using the method offered for that provider, choose a model, and save an agent location. Then assign the saved connection to a persona.

Guides in this section

1. Set up AI providers for your personasConnect an account, choose a model and location, then assign the provider to personas.
2. Connect or repair provider credentialsUse the supported sign-in method and repair expired or rejected credentials.
3. Choose a model and reasoning levelLoad available models and save the reasoning or thinking option your personas need.
4. Choose where your agents runSelect hosted execution or your own cluster and connect private model endpoints.
5. Read usage and set token quotasUnderstand provider allowances, reported token use, and daily token limits.

Your next step

If you are connecting a provider for the first time, follow the Getting Started tutorial before changing advanced settings.

Need a hand? Contact the PersonaStack team.