Neurocaster-Client is an LLM-independent, Angular-based application designed for seamless interaction with multiple LLMs, offering greater control and extensibility.

It supports dynamic chat and tool invocation by integrating with Model Integration Protocol (MIP). The client can work with various LLMs for both conversational AI and executing tools exposed via Spring Boot services or other backends.

• LLM Independence: Works with multiple LLMs for flexible conversational AI and tool invocation.
• Model Integration Protocol (MIP): Seamlessly integrates with services powered by MIP.
• Dynamic Chat: Engage in real-time conversations powered by LLMs.
• Tool Invocation: Execute tools exposed via Spring Boot services or other backends. Call Spring Boot services annotated with @Agent and @Action directly through the client interface.
• Extensibility: Modular architecture allows easy customization and addition of new capabilities.
• Fine-Grained Control: Provides enhanced control over tool execution and LLM interactions.

• Modular and Extensible Architecture: Easy to customize and extend with new functionalities.
• Adaptable to Diverse Workflows: Suitable for various use cases, from conversational AI to complex tool integrations.
• Seamless LLM Interactions: Integrates with multiple LLMs for powerful AI-driven workflows.

Tools4AI is here Tools4AI
Server is here Neurocaster-Server
Client is here Neurocaster-Client

1. Clone the repository:

   git clone https://github.com/yourusername/neurocaster-client.git

2. Navigate to the project directory:

   cd neurocaster-client

3. Install the required dependencies:

   npm install

To run the Angular client, execute the following command:

ng serve

This will start the Angular development server. You can access the client at: http://localhost:4200

💬 Chat with the System Once the client is running, you can initiate a chat with the system. Simply type your query in the chat interface, and the LLM will respond based on the available services exposed via MIP.

🔧 Call Tools The client allows you to invoke tools exposed by the Neurocaster backend. For example, if you have a service like @Action(description = "Customer repair service"), you can call it directly from the client interface.

Type in your query for the tool, such as:

The client will automatically map the request to the appropriate tool method based on the MIP-enabled backend.