Sources in xpander.ai serve as the entry points for AI Agents, initiating tasks and workflows by receiving input from various systems or platforms. Unlike Agentic Interfaces, which enable AI Agents to interact with external systems, Sources specifically handle the entry and starting conditions for AI Agents. They provide the interface through which an AI Agent accesses the Agent Graph System, dynamically invoking graph nodes through function calling.

​
Key Features of Source Nodes:

  • Interface for Graph Nodes: When using the xpander SDK, Source act as the interface for accessing graph nodes in the form of function calling. This allows AI Agents to dynamically retrieve and execute specific tasks within a workflow based on real-time inputs or events. The SDK simplifies the interaction between the source and the agent, ensuring smooth execution of tasks.

  • Native Integration with Platforms: Sources also enable native integration with platforms like Slack, allowing AI Agents to receive inputs and initiate workflows directly from these systems without requiring additional infrastructure management. For example, a Chat Source with Slack integration can listen for messages, commands, or events in Slack, triggering workflows and connecting to the necessary systems.

​
Types of Sources and triggers:

  1. SDK Source Nodes: When using the xpander SDK, The Source provides the interface through which the AI Agent interacts with the Agent Graph System. It facilitates function calling for dynamically generated tasks, allowing the AI Agent to retrieve the necessary graph nodes based on the context of the workflow and the inputs it receives.

  2. Chat Source: provides native integration with Slack or Teams, enabling AI Agents to directly interact with Slack or Teams users through channels, commands, or messages. These source nodes handle the events and pass them to the AI Agent, which can then perform workflows using Slack or Teams without the need for server management, making it easy to deploy intelligent agents for Slack-based workflows.

  3. Webhook Source: These nodes listen for webhook triggers from external systems, allowing AI Agents to initiate tasks based on specific events like form submissions, updates to third-party tools, or new entries in a database.

  4. Task Source Nodes: Add a source node with instructions and a time-based trigger to build Agentic Automation workflows which runs against specific topics in your AI Agent.

TopicsServerless AI Agents