What are Integrations?

​

Introduction

Integrations are connections that enable your AI agents to interact with external platforms, services, and data sources. They serve as the bridge between your agents and the tools they need to perform work on your behalf, from analyzing data in cloud platforms to managing accounting software.

​

Key Concepts

​

What is an Integration?

An integration is a configured connection that:

• Connects to external platforms like Xero, AWS, Azure, and GitHub
• Enables agents to perform actions such as querying databases, managing cloud resources, or accessing documentation
• Provides secure access through authentication credentials or OAuth
• Supports read-only database connections to keep your data safe during analysis
• Ingests data from external sources into the platform for agent analysis

Think of integrations as your agent’s toolkit - they provide the capabilities agents need to accomplish real-world tasks.

​

Types of Integrations

The platform supports two types of integrations:

​

Agent Integrations

Agent Integrations use the Model Context Protocol (MCP) to connect your agents directly to external services and platforms. These integrations enable agents to:

• Query and analyze databases in real-time with read-only access
• Manage cloud infrastructure across AWS, Azure, and other providers
• Interact with business applications like MYOB and Xero
• Access development tools such as GitHub for repository management
• Search documentation using vector-based knowledge bases

Agent Integrations are available to your agents in real-time during conversations and task execution.

​

Data Ingest

Data Ingest integrations use Meltano-Singer taps to bring external data into the platform. Once ingested, this data becomes available for your agents to analyze and work with. Data Ingest supports:

• Technology platform data from Cisco, Microsoft, and other providers
• One-to-One configurations for individual customer accounts
• One-to-Many configurations for managed service providers (MSPs) handling multiple clients
• Scheduled data synchronization to keep information up to date

Data Ingest is ideal when you need historical data analysis or want to work with data offline from its source system.

​

Integration Categories

Integrations are organized into categories to help you find what you need:

​

Authentication Methods

Integrations support multiple authentication approaches:

​

Manual Credentials

For database and API connections, you’ll provide:

• Host URLs and ports
• Usernames and passwords
• Database names or API keys
• Access patterns or permissions

All credentials are encrypted and stored securely.

​

OAuth

For platforms that support OAuth (like MYOB and Xero OAuth2), the authentication process is streamlined:

1. Click “Connect” to begin the OAuth flow
2. You’ll be redirected to the platform’s secure login page
3. Sign in with your credentials
4. Grant permission for Atlastix to access your account
5. You’ll be redirected back and a credential is created automatically
6. Configure any additional settings and save

OAuth credentials can be revoked at any time from your account settings in the external platform.

​

Security and Permissions

​

Read-Only Database Access

When connecting to databases through Agent Integrations, the platform enforces read-only access. This ensures your agents can analyze and query data without the risk of modifying or deleting information.

​

Security Groups

Integrations can be assigned to Security Groups to control which users and agents have access. This allows you to:

• Restrict sensitive integrations to specific teams
• Manage access based on organizational roles
• Maintain compliance with security policies

​

Allowed Patterns

For database integrations, you can define allowed patterns to limit which tables or schemas agents can access, providing an additional layer of security and control.

​

How Agents Use Integrations

When you configure an integration, it becomes available to add to your agents through the agent configuration panel. Once added to an agent:

1. During conversations, the agent can access the integration to retrieve information or perform actions
2. For scheduled tasks, the integration provides the tools needed to complete automated workflows
3. In analysis scenarios, agents can query databases and cloud platforms to generate insights

Agents intelligently determine when to use each integration based on the task at hand and their system instructions.

​

Next Steps