CAITY manager

The execution engine and control center for your AI agents. Scale operations, monitor costs in real time and maintain full control all from a single platform.

End-to-end architecture, designed for production

CAITY follows a streamlined workflow that combines security and efficiency

CAITY is built on a modular, layered architecture.

Skills form the entry point, allowing interaction via web chat, WhatsApp, or direct API access. Incoming requests are then queued to ensure scalability.

At the heart of the system is the application layer, where service logic is developed with CAITY Studio and executed using AI services such as LLMs, alongside databases for state management.

Module

Entry point
The user-facing interface. Available via web chat widget, WhatsApp or direct API.

I/O Queue

Guaranteed scalability
Traffic management queue ensuring scalability and stability even under high load.

App

Agent in execution
Your AI agent executed in a secure, isolated environment.

AI provider

LLM Models
Easy connect to commercial or private LLM providers.

App state

Knowledge base
Persistent memory and knowledge base of the agent, updatable via API.

Session state

Conversation state
Persistent and isolated state for each user session environment

Main features

All the control you need, in one place

Six core features that make CAITY Manager the definitive tool for managing AI agents in production.

Monitoring

Real-time cost and traffic dashboard

Keep track of every euro spent on LLM APIs and every processed request. Dashboard with granularity by agent, by skill, by time period. No surprises at month end.

Conversions

Complete history of every session

Every conversation between user and agent is collected, indexed and made available for analysis, debugging and continuous model improvement. With granular control over data retention.

Deployment

Multi-channel without extra configuration

Deploy your agent on chat widget, WhatsApp or direct API with a few clicks. Each channel has independent style, budget and behavior configurations.

Budget

Budget control per skill

Set a monthly spending cap for each chat configuration. When the limit is reached, the chat automatically suspends. Zero risk of overrun.

Personalization

Fully customizable style and UX

Colors, fonts, status messages, session behavior: everything configurable for each deployment. You can adapt the user experience to your clients' brand without touching the agent's code.

Integration

Open APIs for every system

Open integration layer for seamless connectivity with your business systems. Support for knowledge base updates via API and calls to external services.

See CAITY Manager in action

Book a demo and discover how CAITY Manager simplifies the management of your AI agents in production.