Overview
Live stats for your virtual assistant
—
Active Sessions
—
Total Messages
—
Model
Current Bot Identity
—
—
—
Bot Instructions
Define how your assistant thinks and behaves
System Prompt
This is the core instruction set Claude follows. Be specific about tone, restrictions, and knowledge scope.
Bot Settings
Configure identity, limits, and access
Identity
The Gemini model used for all chat responses.
Limits
Messages kept per session (higher = more memory, more cost)
Max length of each reply
Security
AI Model
Configure your AI provider, model, and API credentials
Connection
Auto-populated for each provider. Edit only if using a proxy or custom gateway.
The model used for all chat responses.
Active Sessions
Live conversation sessions in memory
Sessions
Loading…
Database Memory
Store chat history in PostgreSQL for persistent memory across restarts
Connection Status
Not connected — using in-memory sessions
When enabled, chat history is persisted to your database. Falls back to in-memory if connection fails.
PostgreSQL Credentials
Schema Info
The following table will be created automatically when you connect:
CREATE TABLE chat_sessions (
session_id TEXT PRIMARY KEY,
messages JSONB NOT NULL DEFAULT '[]',
last_active DOUBLE PRECISION NOT NULL
);
Webhook & Integration
Connect this bot to your website
Server URL
The public URL where this server is accessible. Used in the API endpoint and embed snippet below.
Chat API Endpoint
POST
{
"message": "Hello, how can you help me?",
"session_id": "optional-uuid-for-memory"
}
{
"reply": "Hi! I'm here to help...",
"session_id": "20260312-uuid-keep-for-next-message",
"bot_name": "Bell"
}
Website Embed Snippet
Paste this into your website's HTML.
Rate Limits
Current setting: 20 requests per IP per 60 seconds. To adjust, edit app/middleware.py.