Assistant CoreAssistant Core

Create Your First Assistant

Step-by-step guide to creating and configuring an AI assistant.

Overview

An Assistant is your AI agent. It has its own knowledge base, model configuration, voice settings, and can be shared with users or embedded on websites.

Create an Assistant

  1. In the Admin Panel, if you don't have any assistants yet, the system will show a Create New Assistant button. Or navigate directly to: https://admin.assistantcore.com/account/assistant/new

Step 1: Basic Info

Fill in the required fields:

  • Name — Give your assistant a descriptive name (e.g., "Customer Support Bot")
  • Subdomain — Web address for your assistant (auto-generated from name, editable). Example: my-bot.assistantcore.com

Every assistant gets its own unique web address at <name>.assistantcore.com. End users visit this URL to chat with your assistant.

You can replace the default subdomain with your own domain (e.g., chat.yourcompany.com) on the Pro plan or higher. See Custom Domain for setup instructions.

Click Next to proceed to step 2.

Step 2: Description & System Prompt

  • Description — Brief description of what it does (optional)
  • System Prompt — Instructions that define how the assistant behaves

You can use preset templates (Customer Support, Knowledge Base, etc.) or click the Auto Engine ✨ button to let AI generate a system prompt from your name and description.

Click Create to finish. The system will redirect to your new assistant's dashboard.

Models and User Preferences

The create-assistant wizard does not require choosing a model: a new assistant uses the platform defaults. To choose its own models, open Settings → Models in the Admin Panel:

  • Chat model (LLM), and when voice is enabled: Text-to-speech model and voice, Speech-to-text model, Realtime model and voice, and Spoken language.
  • An empty field shows Platform default (model name). Changing a model resets its voice to that model's default.
  • A model that was deactivated after you chose it shows Unavailable; the platform default is used until you pick another.

Widget and API usage is billed to you and always uses this configuration. Signed-in Web App users pay for their own usage and may override it in their own settings. See AI Models for how a model is chosen.

NeedWhere to configure it
Choose the model for text chatUsers choose it in the Web App header. See Chat Model Selection
Choose models and voices for Voice AgentUsers choose them in Settings → Voice in the Web App. See Model and Voice Settings
Enable voice, save recordings, stream response textAdmins configure these in assistant Settings. See Enable Voice
Choose the assistant's modelsAdmins configure them in assistant Settings → Models (above)
Review the model catalog and defaultsSee AI Models

Add Knowledge Base (Optional)

Upload documents so your assistant can answer questions about your content:

  1. Go to Knowledge Base in the sidebar
  2. Click Upload Documents
  3. Drag and drop your files (PDF, DOCX, TXT, XLSX supported)
  4. Wait for processing to complete

See Upload Documents for more details.

Add Users (Optional)

  1. Go to Users in the sidebar
  2. Click Add User
  3. Fill in:
    • Email — Required
    • Full Name — Only needed if creating a new account
    • Role — Choose Owner, Admin, or User
  4. Optional: check Send email notification
  5. Click Add User

If the email already has an account on the platform, the user will be added directly. If not, the system will create a new account and send an invitation.

User limits

User limits can change with the current SaaS plan configuration. See Plans & Pricing for the latest limits.

Roles & Permissions

ActionOwnerAdminUser
Chat with assistant✅✅✅
View conversation history✅✅Own only
Upload documents✅✅❌
Change assistant settings✅✅❌
Manage users✅✅❌
Manage subscription✅❌❌
Delete assistant✅❌❌

Maximum users and resource limits depend on the owner's plan. See Plans & Pricing for the latest limits, and Credits & Usage to understand credit consumption.

What's Next?

On this page