10. Plugins & Extensions

MindLogic provides a lightweight and powerful Plugin System. In addition to Entity Scripts, you can expand MindLogic's capabilities by installing plugins to connect to external APIs, process complex text, or invoke Large Language Models (LLMs).

Plugin Applications

Plugins are primarily used to encapsulate complex network requests, data processing, and third-party service integrations. Typical scenarios include:

• API Requests: Interact with various internet services (e.g., getting weather, stocks, exchange rates).
• AI Model Integration: Call various Large Language Models (OpenAI, DeepSeek, etc.) for automated analysis.
• Information Scraping & Extraction: Extract text from web pages, PDFs, and other external documents for reasoning.
• Toolchain Integration: Invoke search engines or specific domain-professional tools.

Official Plugins Download List

You can directly download our official plugins and import them through the Tools -> Plugin Center inside the app.

(Note: The Test plugin is for internal development only and is not listed here)

Plugin Usage Guide

1. Install Plugins: Download the .mlplugin files above. Open MindLogic, go to the top menu bar Tools -> Plugin Center, click the Import icon in the top right corner, and select the downloaded file.

2. Global Configuration: Some plugins (like the OpenAI LLM plugin) may require an API Key or other global parameters. Select the installed plugin in the "Plugin Center", click Edit Global Config, fill in your Key, and save. These configurations apply to all nodes using this plugin.

3. Using in Nodes: Select any node on the canvas. At the bottom of the right-hand Inspector panel, expand the Plugin section and select the required plugin from the dropdown list. Once selected, specific input fields for that plugin will appear (e.g., Prompt, Target URL). You can input static text or use the {} syntax to reference data from upstream nodes (e.g., { node.inputs['City'] }).

4. Viewing Execution Results: When the graph computation flows to the node and all inputs are ready, the plugin executes automatically in the background. Upon success, its result is injected into the node's properties, allowing you to use it in downstream nodes.

Execution Order & Data Flow Logic

To maximize the power of plugins and scripts, it is crucial to understand the exact execution order of a MindLogic node. The computation flows through three main stages:

1. Math & Formula Engine (First): The system first pulls in all Inputs propagated from upstream nodes, and evaluates all defined "Formulas" in the Inspector. At this stage, node.inputs and node.outputs are assigned their initial, statically calculated values.
2. Plugin Script (Second): If a plugin is configured for the node, its underlying code snippet executes immediately after the formulas. Plugins typically read data via { variable } syntax, make network/AI requests via system APIs, and assign the retrieved results directly into node.outputs['Property'] or modify the node's title (node.title).
3. Node Custom Script (Last): The custom "Script" at the very bottom of the Inspector panel has the final say. Because it runs consecutively after the plugin script, you can directly access the plugin's output results here! For example, if an AI plugin writes its result to node.outputs['Summary'], your custom script can read it:

   let result = node.outputs['Summary'];
   if (result.includes("Error")) {
       node.title = "Status: Failed";
   }
   

   Finally, the final outputs variables of the current node are automatically passed along the edge connections to become the inputs for the next downstream node.

Writing Custom Plugins

Besides official plugins, you can easily write your own! MindLogic provides a powerful built-in Plugin Studio, so you don't need to manually write JSON files or zip archives. You can visually create and edit plugins directly within the app.

Creating and Editing

1. Open Tools -> Plugin Center.
2. Click the Create Plugin button in the top right corner to open the Plugin Studio.
3. In the left panel, you can configure the plugin's metadata (Name, Author, etc.) and define Global Inputs (like API Keys) as well as Node Inputs (UI form fields presented to the user).
4. In the Script Template on the right panel, you can write the specific execution logic using JavaScript.

Debugging & Development

MindLogic provides a secure Sandbox Testing environment:

• After writing your code in the Plugin Studio, you can directly click the Test button.
• In the sandbox window that appears, enter mock test data and click Run to view the execution logs and results.
• This allows you to test and refine the plugin code in isolation without affecting your canvas document. Once done, click Save & Install to apply the plugin locally.