This article provides a comprehensive guide to understanding and utilizing apps within AgentRunner, an agent builder platform using a node editor. Whether you're a new user or looking to expand your knowledge, this resource will help you navigate the app ecosystem and integrate external software seamlessly into your agents. AgentRunner empowers you to connect various software applications through APIs, enabling automated data exchange and streamlined processes. This guide covers everything from finding and connecting to existing apps to creating your own custom integrations, ensuring you can leverage the full potential of AgentRunner's app functionality.
Exploring the Apps Page
The Apps page in AgentRunner serves as a central hub for discovering and managing pre-built software integrations accessible via APIs (Application Programming Interfaces). An API allows you to call, run, and automate agents programmatically, facilitating data exchange without manual intervention. In practical terms, connecting through an API means you can link AgentRunner with other software to automatically exchange data and execute functions, such as retrieving data, triggering actions, or synchronizing information between systems. These apps are listed in the left-hand menu, offering a streamlined way to enhance your agents.
Navigating the Apps Interface
When you access the Apps page, you will find a comprehensive list of available apps within your workspace. These apps are categorized into three distinct types:
Private Apps: These are created by your team and are exclusive to your workspace.
Public Apps: These are built either within your workspace or by other users and made available for general use.
LLM Apps: Apps representing Large Language Models (LLMs), like OpenAI, Gemini, or Anthropic. LLMs are AI models that can be integrated to add AI capabilities to the agents.
Connecting and Managing Apps
Once you've located an app that meets your needs, you can access its dedicated page to view additional details and connect it to your AgentRunner workflow. On the app page, you will find a brief description of the app, along with a list of available endpoints. Endpoints are specific functions or actions that the app can perform, which you can then utilize within your agents. To establish a connection, you'll need to provide your API key for the respective application.
Connecting via API Key
Currently, AgentRunner supports API key connections, with OAuth authentication coming soon. To connect an app using an API key, follow these steps:
On the app page, click the "Connect with API" button.
A modal will appear, prompting you to enter your API key.
Provide a name for your connection. This is important if you plan to connect multiple API keys for the same app, as it allows you to easily differentiate between them.
Enter your API key in the designated field.
Tip: Ensure that you accurately enter your API key, as an incorrect or invalid key may result in API errors when running the node.
Managing Connected Apps
Once the API is connected, it will be displayed on the app page, showing its name, status (connected), creation date, and options to rename or disconnect it. You can connect a limitless number of apps. Expected Outcome: After successfully connecting the API key, you should see a success message and the connection details displayed on the app page.
Creating Custom Apps
If you can't find a specific app that meets your requirements, AgentRunner offers the flexibility to create your own custom apps. To initiate this process, click the "Create App" button located in the upper right corner of the Apps page. This will present you with two options: requesting a new app or creating one yourself.
Requesting a New App Integration
If you prefer to have AgentRunner handle the integration, you can request a new app by providing the necessary details:
Click the "Request a New App" option.
Provide the integration coordinates, including the website URL, API documentation (if available), and contact information for the company you want to integrate with.
Click the "Send Request" button.
AgentRunner will then evaluate your request and strive to integrate with the specified application. By requesting a new app, you are essentially suggesting an integration to the AgentRunner team, who will assess its feasibility and prioritize it based on user demand and technical considerations. It's a great way to expand the app ecosystem and get the integrations you need.
Creating Your Own App
If you have the technical expertise, you can create your own app from specifications. This option requires some familiarity with API structures, especially with the OpenAPI schema. If you're not familiar with OpenAPI, you may not be able to complete this task. Here's how to create your own app:
Click the "Create Your Own App" option.
Upload an image for your app.
Enter a name for your app.
Select a color for your app's label.
Provide the website URL for the app.
Enter a short and long description for your app.
Upload an OpenAPI schema that specifies the app's endpoints and how to use them.
Save your app.
Tip: All apps, including private ones, are stored and accessible under the Apps section in AgentRunner.
Utilizing AI-Focused Apps
AgentRunner features AI models, such as OpenAI, Gemini, and Anthropic, which are slightly different from generic apps. These models have a special AI focus as an editor in the node editor. AI models are accessible through special nodes in the node editor, designed for high-level prompt engineering. These specialized nodes enable you to create sophisticated AI agents and build innovative AI applications. The node editor is the place where you can create agents and each workflow is managed in the agent editor. AI nodes like OpenAI, Gemini, and Anthropic have dedicated nodes in the node editor that are specifically designed for advanced prompt engineering, empowering you to build cutting-edge AI applications.
Working with AI Nodes
To effectively use AI-focused apps, consider the following:
Understand Prompt Engineering: Learn how to craft effective prompts to guide AI models.
Explore Node Editor Features: Familiarize yourself with the specific features of the node editor designed for AI models.
Experiment with Different Models: Test different AI models to determine which one best suits your needs.
By understanding and utilizing the app ecosystem within AgentRunner, you can significantly enhance your agents, automate tasks, and unlock new possibilities for your projects. Whether you leverage existing apps, request new integrations, or create your own custom solutions, AgentRunner empowers you to connect and integrate with the tools and services you need to succeed.
