MakerAI Suite: Advanced AI Components for Delphi

📌 Project Description

The MakerAI Suite is a comprehensive set of Delphi components designed to seamlessly integrate artificial intelligence into your applications. With support for state-of-the-art models and functionalities, the suite includes tools for natural language processing, audio transcription, image generation, task orchestration, and retrieval-augmented generation (RAG).

Core Components

• Chat: Integration with models like Anthropic, Gemini, Grok, Groq, Mistral, Ollama, and OpenAI.
• Embeddings: Vector-based text representations using Grok, Groq, Mistral, Ollama, and OpenAI.
• Audio: Powered by Whisper (OpenAI), for transcription and translation.
• RAG: Retrieval-augmented generation using all supported models.
• Graph: Task orchestration via TAiGraph, enabling visual workflow creation.

🌟 Key Features

TAiChat

• Broad Compatibility: Supports multiple AI models like GPT-4, Claude, Mistral, and more.
• File Attachments: Processes images, audios, and other media inputs.
• Async Mode: Real-time feedback for improved user experiences.
• Tool Integration: Enables interaction with external tools for queries and tasks.

TAiAudio

• Transcription: Converts audio into text.
• Translation: Translates audio content between languages.
• Voice Synthesis: Generates spoken audio from text.

TAiDalle

• Image Generation: Creates images from text descriptions.
• Image Editing: Modifies existing images using masks.
• Variations: Produces alternate versions of an image.

TAiGraph

• Visual Task Orchestration: Simplifies workflow design using graph-based structures.
• Modular Design: Facilitates the integration of AI components and external tools.
• Contextual Workflows: Builds dynamic systems that adapt to changing contexts.

RAG Integration

• Contextual Queries: Combines language models with context retrieved from vector databases.
• Database Support: Works with any vector database, like Pinecone or PostgreSQL (pg_vector) and in-memory embeddings.
• Scalability: Handles large datasets for advanced AI-powered systems.

🎯 Use Cases

🌐 TAiChat

1. Virtual Assistants: Manages complex queries and provides contextual support.
2. Sentiment Analysis: Detects tones in social media or survey data.
3. Content Generation: Summarizes, generates reports, or writes articles.

🎧 TAiAudio

1. Automatic Subtitling: Creates subtitles for videos.
2. Voice Assistants: Enables voice commands for chatbots and applications.
3. Meeting Documentation: Transcribes conferences or interviews.

🎨 TAiDalle

1. Visual Design: Creates illustrations from descriptions.
2. Prototyping: Generates quick visual concepts.
3. Creative Editing: Enhances images using AI.

🧩 TAiGraph

1. Task Automation: Builds workflows for business processes.
2. AI-Enhanced Operations: Integrates AI models into dynamic, adaptive pipelines.
3. Visual System Design: Creates modular systems with intuitive graph interfaces.

🔍 RAG

1. Semantic Search: Retrieves precise information from large datasets.
2. Knowledge-Based Systems: Enhances AI responses with specific contextual knowledge.
3. Custom AI Assistants: Builds powerful tools for industries like healthcare or finance.

📚 Examples

🛠️ TAiChat

var
  Chat: TAiChat;
begin
  Chat := TAiChat.Create(nil);
  try
    Chat.ApiKey := 'your-api-key';
    Chat.Model := 'gpt-4';
    Chat.AddMessage('What is the capital of France?', 'user');
    ShowMessage(Chat.Run);
  finally
    Chat.Free;
  end;
end;

🧩 TAiChat with images

var
     Res: String;
     MediaFile: TAiMediaFile;
   begin
     MediaFile := TAiMediaFile.Create;
     MediaFile.LoadFromFile('ruta/del/archivo.jpg');
     Res := Chat.AddMessageAndRun('Describe esta imagen', 'user', [MediaFile]);
     ShowMessage(Res);
     MediaFile.Free;
   end;

🔍 TAiChat with voice files

Var
  Res: String;
  MediaFile: TAiMediaFile;
  Msg: TAiChatMessage;
  FileName: String;
begin

  MediaFile := TAiMediaFile.Create;
  MediaFile.LoadFromFile('c:\temp\prompt.wav');

  Try
    Msg := AiOpenChat1.AddMessageAndRunMsg(MemoPrompt.Lines.Text, 'user', [MediaFile]);
  Finally
    FreeAndNil(MediaFile);
  End;

  MemoResponse.Lines.Text := Msg.Content;

  If (Msg.MediaFiles.Count > 0) and (Assigned(Msg.MediaFiles[0].Content)) then
  Begin
    FileName := 'c:\temp\respuesta3' + Cons.ToString + '.wav';
    Msg.MediaFiles[0].Content.Position := 0;
    Msg.MediaFiles[0].Content.SaveToFile(FileName);

    Try
      MediaPlayer1.FileName := FileName;
      MediaPlayer1.Play;
    Finally
    End;
  End;

🛠️ Setup

Requirements

1. Delphi 11 or higher.
2. API keys for supported models (e.g., OpenAI, Anthropic).
3. Dependencies:
   • System.Net.HttpClient
   • System.JSON
   • REST.Client

Installation

1. Clone this repository.
2. Configure API keys in the component properties (e.g., ApiKey).
3. Follow the examples to integrate components into your Delphi project.

📜 License

This project is licensed under the MIT License.

👤 Author

Gustavo Enríquez

• LinkedIn: Profile
• YouTube: Channel
• GitHub: Repository

Want to contribute? Feel free to fork and suggest improvements!