Seraph AI System

Seraph is an advanced AI assistant system designed to provide intelligent, context-aware interactions through a CLI interface. It features a sophisticated cognitive architecture with memory persistence, function execution capabilities, and middleware support.

Features

• Persistent Memory: Long-term storage and retrieval of information using semantic search
• Cognitive Functions: Extensible system of tools and functions that Seraph can use to help users
• Middleware System: Pluggable middleware architecture for customizing behavior and adding capabilities
• CLI Interface: Clean, user-friendly command-line interface with formatted output
• Conversation History: Maintains context across sessions with conversation persistence

Installation

1. Clone the repository:

git clone https://github.com/oneirocom/seraph.git
cd seraph

2. Install dependencies:

npm install

3. Set up environment variables:

cp .env.example .env

Add your API keys to the .env file:

OPENAI_API_KEY=your_openai_key_here
ANTHROPIC_API_KEY=your_anthropic_key_here

4. Build the project:

npm run build

Usage

Start Seraph:

npm start

Available Commands

• !help - Display available commands
• !vim - Open text editor for multi-line input
• !prompt - Display current system prompt
• !messages - Show conversation history
• !clearLastMessage - Remove last message
• !memory-diagnostics - Run diagnostics on memory system

Type exit to end the conversation.

Architecture

Seraph is built with a modular architecture consisting of several key components:

Core Components

• SeraphCore: Main system orchestrator
• ConversationManager: Handles conversation state and persistence
• CognitiveFunctionExecutor: Manages execution of cognitive functions
• MiddlewareManager: Handles middleware registration and execution
• ResponseParser: Processes and parses AI responses
• MemoryStore: Manages semantic memory storage and retrieval

Memory System

Seraph uses a sophisticated memory system with different types:

• code: Stores generated code and related information
• context: Stores contextual information
• reflection: Stores self-reflection and performance metrics
• facts: Stores facts and general knowledge
• system: Stores system-related information

Cognitive Functions

Seraph comes with built-in cognitive functions:

• Memory Storage and Retrieval
• Bash Command Execution
• Custom function support

Development

Project Structure

src/
├── lib/
│   ├── cognitive_functions/    # Cognitive function implementations
│   ├── middleware/            # Middleware implementations
│   ├── seraphCore.ts         # Core system implementation
│   ├── conversation_manager.ts # Conversation handling
│   ├── response_parser.ts    # Response parsing logic
│   └── seraphCLI.ts         # CLI interface

Adding New Functions

1. Create a new class extending BaseCognitiveFunction
2. Implement the required methods
3. Register the function in src/lib/cli.ts

Example:

class NewFunction extends BaseCognitiveFunction {
  constructor(seraph: SeraphCore) {
    super({
      name: "newFunction",
      description: "Description of the function",
      parameters: {
        param1: {
          type: "string",
          description: "Parameter description",
        },
      },
    });
  }

  async execute(args: Record<string, any>): Promise<string> {
    // Implementation
  }
}

Adding New Middleware

1. Create a new class implementing IMiddleware
2. Implement the required methods
3. Register the middleware in src/lib/cli.ts

Example:

class NewMiddleware implements IMiddleware {
  name = "newMiddleware";
  schema = z.object({
    // Schema definition
  });

  async run(response: string): Promise<string> {
    // Implementation
  }
}

Contributing

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Create a Pull Request

License

MIT License - See LICENSE file for details