Aide - Personal AI Assistant

Overview

Aide is a .NET-native AI agent runtime and personal assistant that helps manage your digital communications and tasks through natural language interaction. Designed to support the Model Context Protocol (MCP), Aide provides a high-performance, extensible platform for AI-powered automation with first-class .NET ecosystem integration.

Relationship to Moltbot/OpenClaw

Aide shares similar goals with Moltbot/OpenClaw, the popular self-hosted AI assistant. Both projects aim to provide:

Self-hosted, privacy-first personal AI assistants
Plugin/capability-based extensibility
Multi-service integration (email, Discord, etc.)
Action-oriented intelligence that "actually does things"

Aide's Differentiation:

Native .NET Performance - Built on .NET 10 for maximum performance and type safety
.NET Ecosystem Integration - First-class support for .NET libraries, Active Directory, Windows services
Enterprise Focus - Advanced security, compliance, audit logging, and enterprise deployment
MCP + C# Plugins - Supports both Model Context Protocol and native C# plugins
Advanced Memory Architecture - PostgreSQL + Vector DB for sophisticated context management
.NET Developer Experience - Familiar tooling and patterns for C# developers

Aide is designed to be the best-in-class AI agent runtime for the .NET ecosystem while leveraging the broader MCP community.

Architecture

Aide is built as a capability runtime rather than a monolithic application:

Core Components

Runtime Engine - .NET-based orchestration and execution environment
LLM Gateway - Multi-provider support (Claude, GPT, Gemini, local models)
Capability System - Dual plugin architecture:
- Native C# Plugins - High-performance compiled capabilities for .NET developers (MVP)
- MCP Capabilities - Leverage 100+ existing Model Context Protocol integrations (Planned)
Background Task Queue - Execute long-running operations asynchronously
- In-memory queue for MVP
- Support for scheduled/recurring tasks (Post-MVP)
- Task status tracking and cancellation
Memory Service - Advanced context persistence using PostgreSQL + Vector DB
Security Layer - Plugin sandboxing, permission model, and audit logging
Audit System - Comprehensive logging of all LLM actions, tool calls, and results
- Track what AI decided and why
- Monitor costs (token usage, API calls)
- Compliance for financial/trading operations
- Debugging and transparency
- Immutable, searchable audit trail

Capability Examples

Communication capabilities (via MCP or native plugins):

Email Integration - Gmail, Outlook, IMAP/SMTP
Discord Integration - Server monitoring, message management
Slack, Teams, etc. - Additional messaging platforms

Intelligence capabilities:

Task Management - Create, track, and complete tasks
Background Operations - Long-running tasks, scheduled monitoring, async processing
Smart Filtering - Spam detection and priority routing
Context Analysis - Message categorization and insights
Automation - Custom workflows and batch operations

Interaction capabilities:

Natural Language Chat - Text-based interface
Voice Input - Speech-to-text (OpenAI Whisper, Azure Speech, etc.)
Voice Output - AI-powered text-to-speech (ElevenLabs, OpenAI TTS, Azure Neural Voices)
Hands-free Mode - Continuous voice conversation
Desktop Integration - Windows native features

Technology Stack

Core Runtime

.NET 10 - High-performance runtime
ASP.NET Core - Web API and services
C# 13 - Utilizing the most modern C# syntax and language features available
- Primary constructors
- Collection expressions
- Required members
- File-scoped types
- Raw string literals
- Pattern matching enhancements

Capability System

Model Context Protocol (MCP) - Standard protocol for AI tool integration
AssemblyLoadContext - Dynamic plugin loading and isolation
Native C# Plugins - Compiled capabilities using ICapability interface

LLM Integration (Multi-Provider)

Anthropic Claude - Claude Opus, Sonnet, Haiku
OpenAI - GPT-4, GPT-4 Turbo, GPT-3.5
Google Gemini - Gemini Pro, Ultra
Local Models - Ollama, LM Studio
Custom - Any OpenAI-compatible API

Data Layer

PostgreSQL - Structured data, tasks, audit logs
Vector Database - Pinecone, Weaviate, or Chroma for semantic memory
Redis - Caching, sessions, real-time state

Voice Integration (Planned)

Text-to-Speech - ElevenLabs (primary), OpenAI TTS, Azure Neural Voices, Google Cloud TTS
Speech-to-Text - OpenAI Whisper, Azure Speech Services, Google Speech-to-Text
Multi-provider support - Abstraction layer for switching voice providers

Frontend

.NET MAUI Blazor Hybrid - Cross-platform mobile and desktop (iOS, Android, Windows, macOS)
Blazor - Web-based UI components in native app shell
MudBlazor - Material Design component library for styling and UI components
Tailwind CSS - Utility-first CSS for positioning, layouts, and animations
Platform Integration - Native features per platform (microphone, audio playback)

Project Structure (Planned)

Aide/
├── src/
│   ├── Aide.Core/                    # Core runtime and abstractions
│   │   ├── Abstractions/             # ICapability, ILlmProvider, etc.
│   │   ├── Runtime/                  # Plugin loader, orchestrator
│   │   ├── Services/                 # Memory, LLM gateway, security
│   │   └── Mcp/                      # Model Context Protocol client
│   │
│   ├── Aide.Api/                     # ASP.NET Core Web API
│   │   ├── Controllers/
│   │   ├── SignalR/                  # Real-time communication
│   │   └── Program.cs
│   │
│   ├── Aide.Capabilities/            # Built-in native capabilities
│   │   ├── Email/
│   │   ├── Discord/
│   │   ├── Tasks/
│   │   └── FileSystem/
│   │
│   ├── Aide.Sdk/                     # SDK for plugin developers
│   │   ├── CapabilityBase.cs
│   │   ├── Attributes/
│   │   └── Testing/
│   │
│   ├── Aide.Voice/                  # Voice interaction services (planned)
│   │   ├── Abstractions/
│   │   │   ├── ITextToSpeechProvider.cs
│   │   │   └── ISpeechToTextProvider.cs
│   │   ├── Providers/
│   │   │   ├── ElevenLabsProvider.cs
│   │   │   ├── OpenAiTtsProvider.cs
│   │   │   └── WhisperProvider.cs
│   │   └── VoiceOrchestrator.cs
│   │
│   └── Aide.Ui/                     # .NET MAUI Blazor Hybrid UI
│       ├── Platforms/               # iOS, Android, Windows, macOS
│       ├── Components/              # Blazor components
│       │   ├── Pages/               # Chat, Settings, Capabilities
│       │   ├── Shared/              # Layout, NavMenu
│       │   └── UI/                  # Reusable UI components
│       ├── wwwroot/                 # Static assets, Tailwind CSS
│       │   ├── css/
│       │   ├── js/
│       │   └── index.html
│       ├── Services/                # API client, state management
│       └── MauiProgram.cs
│
├── plugins/                          # External plugins (MCP + C# DLLs)
│   ├── mcp-servers/                 # MCP server configurations
│   └── custom/                      # Custom C# plugin DLLs
│
├── docs/                            # Documentation
│   ├── architecture.md
│   ├── plugin-development.md
│   └── mcp-integration.md
│
└── tests/                           # Unit and integration tests

Development Roadmap

MVP (Phase 1-7)

Goal: Working AI agent with Claude integration, sample capabilities, and audit logging

Estimated Time: 20-25 hours (experienced .NET developer), 30-40 hours (learning)

Post-MVP Enhancements

Phase 8.1: Additional LLM Providers

OpenAI (GPT-4) provider
Google Gemini provider
Ollama (local models) provider

Phase 8.2: MCP Integration

MCP protocol client implementation
MCP server discovery and lifecycle management
Bridge MCP capabilities to ICapability interface
Test with existing MCP servers

Phase 8.3: Plugin Loading

AssemblyLoadContext-based dynamic plugin loader
Load C# DLLs from directories
Plugin manifest files and hot reload

Phase 8.4: Real Capabilities

Phase 8.5: Voice Interaction

Voice provider abstraction
ElevenLabs TTS integration
OpenAI Whisper STT integration
Voice UI in MAUI app

Phase 8.6: Advanced Task Scheduling

Persistent task storage (PostgreSQL)
Cron-based scheduling (Hangfire or Quartz.NET)
Recurring task management
Task history and results
UI for managing scheduled tasks

Phase 8.7: Advanced Features

PostgreSQL + EF Core integration
Vector database for semantic memory
Multi-user support and authentication
Plugin marketplace

Core Principles

Self-hosted - Run entirely on your own infrastructure
Privacy-first - No data sharing with third parties
Extensible - MCP + native C# plugin support
Multi-LLM - Not locked into a single provider
Type-safe - Leverage C# and .NET's type system
Secure - Plugin sandboxing, encrypted credentials, audit logging
Enterprise-ready - Active Directory, compliance, deployment tools

Getting Started

Prerequisites

.NET 10 SDK - Download here
Bun - Install from bun.sh (for Tailwind CSS builds)
Anthropic API Key - Get from console.anthropic.com
IDE - Visual Studio 2022, VS Code, or Rider

Development Setup

1. Clone the Repository

git clone https://github.com/saib-inc/aide
cd aide

2. Configure API Key

Create a configuration file at ~/.aide/appsettings.json:

{
  "Aide": {
    "Llm": {
      "Providers": {
        "Claude": {
          "ApiKey": "your-anthropic-api-key-here"
        }
      }
    }
  }
}

Note: This file is in your home directory and won't be committed to git. You can also set the API key via environment variable: ANTHROPIC_API_KEY

3. Install UI Dependencies

cd src/Aide.Ui/wwwroot
bun install
cd ../../..

Running the Project

Option A: Run Both API and UI (Recommended)

Terminal 1 - Start the API:

dotnet run --project src/Aide.Api/Aide.Api.csproj

The API will start at http://localhost:5009

Terminal 2 - Start the UI:

dotnet run --project src/Aide.Ui/Aide.Ui.csproj -f net10.0-maccatalyst

Note: Tailwind CSS automatically rebuilds when you build the UI project (configured in Aide.Ui.csproj)

Option B: Run API Only (for testing)

dotnet run --project src/Aide.Api/Aide.Api.csproj

Test the API:

# Get capabilities
curl http://localhost:5009/api/capabilities

# Send a chat message
curl -X POST http://localhost:5009/api/chat \
  -H "Content-Type: application/json" \
  -d '{"message": "Hello! Can you call hello_world?"}'

Option C: Build and Run for Production

# Build everything
dotnet build

# Run API
dotnet run --project src/Aide.Api/Aide.Api.csproj -c Release

# Run UI
dotnet run --project src/Aide.Ui/Aide.Ui.csproj -f net10.0-maccatalyst -c Release

Platform-Specific UI Builds

The UI is built with .NET MAUI and supports multiple platforms:

macOS (Catalyst):

dotnet run --project src/Aide.Ui/Aide.Ui.csproj -f net10.0-maccatalyst

iOS:

dotnet run --project src/Aide.Ui/Aide.Ui.csproj -f net10.0-ios

Android:

dotnet run --project src/Aide.Ui/Aide.Ui.csproj -f net10.0-android

Windows:

dotnet run --project src/Aide.Ui/Aide.Ui.csproj -f net10.0-windows10.0.19041.0

Development Workflow

API Development:
- API auto-reloads with dotnet watch run --project src/Aide.Api/Aide.Api.csproj
- Add capabilities in src/Aide.Capabilities/
- Check logs for LLM tool calls and responses
UI Development:
- Tailwind CSS rebuilds automatically on project build
- Manual Tailwind rebuild: cd src/Aide.Ui/wwwroot && bun run build:styles
- UI uses MudBlazor for components + Tailwind for layout
- Hot reload: dotnet watch run --project src/Aide.Ui/Aide.Ui.csproj -f net10.0-maccatalyst
Testing the Integration:
- Start both API and UI
- Open the UI app
- Try the sample prompts or type a message
- Check API terminal for capability execution logs

Troubleshooting

"Failed to load capabilities" in UI:

Make sure the API is running at http://localhost:5009
Check the API terminal for errors

"500 Internal Server Error" from API:

Verify your API key is configured in ~/.aide/appsettings.json
Check the API terminal for detailed error messages

Tailwind styles not applying:

Run cd src/Aide.Ui/wwwroot && bun run build:styles
Make sure Bun is installed and in your PATH

UI won't start:

Make sure .NET 10 SDK is installed: dotnet --version
For macOS: Xcode Command Line Tools required
For Android: Android SDK required
For iOS: Xcode and provisioning profile required

Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines on:

Conventional commits
Pull request process
Code standards
Development workflow

Repository: https://github.com/saib-inc/aide

License

To be determined

Development Status

🚧 Planning Phase - Architecture and technology decisions in progress

Name		Name	Last commit message	Last commit date
Latest commit History 8 Commits
src		src
tests/Aide.Core.Tests		tests/Aide.Core.Tests
.gitignore		.gitignore
Aide.slnx		Aide.slnx
CONTRIBUTING.md		CONTRIBUTING.md
PLAN.md		PLAN.md
README.md		README.md

SAIB-Inc/aide

Folders and files

Latest commit

History

Repository files navigation

Aide - Personal AI Assistant

Overview

Relationship to Moltbot/OpenClaw

Architecture

Core Components

Capability Examples

Technology Stack

Core Runtime

Capability System

LLM Integration (Multi-Provider)

Data Layer

Voice Integration (Planned)

Frontend

Project Structure (Planned)

Development Roadmap

MVP (Phase 1-7)

Post-MVP Enhancements

Core Principles

Getting Started

Prerequisites

Development Setup

1. Clone the Repository

2. Configure API Key

3. Install UI Dependencies

Running the Project

Option A: Run Both API and UI (Recommended)

Option B: Run API Only (for testing)

Option C: Build and Run for Production

Platform-Specific UI Builds

Development Workflow

Troubleshooting

Contributing

License

Development Status

About

Resources

Contributing

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Contributors 2

Uh oh!

Languages

Packages