Coding Assistants

A multi-agent orchestration desktop application that coordinates multiple LLM-powered agents to collaboratively solve software engineering tasks. Configure agent roles (Planner, Developer, Reviewer, etc.), assign different LLM providers to each, and watch them work together with inter-agent communication, user interaction, and real-time streaming output.

---

Features

• Multi-Agent Orchestration -- Define multiple agent roles with independent configurations and watch them execute tasks sequentially, passing context between phases.
• Multi-Provider LLM Support -- Connect to OpenCode Zen, Google Gemini, Anthropic Claude, OpenAI, GitHub Copilot, Ollama, and LM Studio from a single interface.
• Real-Time Streaming -- Agent thoughts and responses stream to the UI line-by-line as they are generated.
• Inter-Agent Communication -- Agents can request input from other agents via [[ASK_AGENT:RoleName]] markers, with user authorization.
• User-in-the-Loop -- Agents can pause and ask the user questions via [[ASK_USER]] markers, displayed as modal dialogs.
• Workspace-Aware -- Select a project directory and agents operate within it, reading/writing files and loading custom prompts, rules, and workflows from .agent/.
• Remote Control -- Start a TCP server on port 5555 and control the desktop app from an Android companion app over your local network.
• MCP Integration -- Configure Model Context Protocol servers (sequential-thinking, filesystem, memory) for enhanced agent capabilities.
• Glass-Morphism UI -- Dark theme with frosted glass effects, gradient buttons, and colored agent badges.

Screenshots

Coming soon

Quick Start

Prerequisites

• Node.js (LTS recommended)
• Rust (stable toolchain)
• Tauri CLI prerequisites for your platform
• At least one LLM provider:
  • Ollama for local models
  • API keys for cloud providers (OpenAI, Anthropic, Google, etc.)

Install & Run

# Clone the repository
git clone https://github.com/your-username/Coding-Assistants.git
cd Coding-Assistants

# Install frontend dependencies
npm install

# Run in development mode (launches both frontend and Tauri backend)
npm run tauri dev

Configure API Keys

Copy the environment template and add your keys:

cp env/vars.env env/.env
# Edit env/.env with your API keys

Build for Production

npm run tauri build

The bundled application will be output to src-tauri/target/release/bundle/.

Architecture Overview

┌─────────────────────────────────────────────────┐
│                  React Frontend                  │
│  ┌───────────┐ ┌──────────┐ ┌────────────────┐  │
│  │ Role Mgmt │ │ Task Exec│ │ Event Viewer   │  │
│  └─────┬─────┘ └────┬─────┘ └───────┬────────┘  │
│        │            │               │            │
│        └────────────┼───────────────┘            │
│                     │ invoke / listen             │
├─────────────────────┼───────────────────────────-┤
│                     │ Tauri IPC                   │
├─────────────────────┼────────────────────────────┤
│                  Rust Backend                     │
│  ┌──────────┐ ┌─────┴─────┐ ┌──────────────┐    │
│  │ AgentSys │ │ LLMClient │ │  TCP Server   │    │
│  └──────────┘ └───────────┘ └──────────────-┘    │
│        │            │               │            │
│        ▼            ▼               ▼            │
│   File Tools    LLM APIs     Android App         │
└─────────────────────────────────────────────────-┘

See ARCHITECTURE.md for a detailed breakdown.

Supported LLM Providers

Provider	Type	Models
OpenCode Zen	Cloud	Auto-detected via CLI
Google Gemini	Cloud	Gemini Pro, Flash, etc.
Anthropic	Cloud	Claude 3/3.5/4 family
OpenAI	Cloud	GPT-4o, GPT-4, etc.
GitHub Copilot	Cloud	Copilot models
Ollama	Local	Any pulled model
LM Studio	Local	Any loaded model

Android Companion App

The android/ directory contains a Kotlin/Jetpack Compose app for remote controlling the desktop application over TCP/IP. See the Android README for setup instructions.

Documentation

Document	Description
AGENTS.md	Governance and workflow for AI assistants
ARCHITECTURE.md	System design and data flow
DEPENDENCIES.md	Dependency inventory and rationale
DEVELOPMENT.md	Developer setup and workflow guide
TESTING.md	Testing strategy and instructions
SECURITY.md	Security policy and guidelines
CONTRIBUTING.md	Contribution workflow
TUTORIAL.md	Step-by-step usage tutorial
TROUBLESHOOTING.md	Common issues and solutions
ROADMAP.md	Project roadmap and planned features

Recommended IDE Setup

• VS Code + Tauri Extension + rust-analyzer

License

This project is licensed under the GNU Affero General Public License v3.0.