Skip to content

STiFLeR7/personal-agent-os

Repository files navigation

Dex Logo

Dex | Personal Discord Cognitive Operator

v1.2 Hermes Render

An elite, proactive cognitive control plane living in your private Discord.
Dex is a dual-core AI operator that manages your digital life through planning, task management, and multi-channel delivery.

Dex Banner


📖 Overview

Dex is a Cognitive Operating System designed for high-performance individuals. It transforms Discord from a chat app into a command center. Unlike traditional bots, Dex uses a "Hermes-style" delivery system—it is proactive, branded, and deeply integrated into your server's workflow.

By leveraging Gemini 2.0 Flash for reasoning and Groq (Llama 3.3) as a high-speed fallback, Dex ensures your personal assistant is always available, always dynamic, and always secure.

Why Dex?

  • Proactive Intelligence: Not just a chatbot; Dex initiates tasks and provides daily briefings.
  • Dual-Core Brain: High-fidelity reasoning with Gemini; ultra-fast fallback with Groq.
  • Branded Interface: Native Discord embeds with custom branding and "Hermes" messenger layer.
  • Full Interactivity: Talk to Dex anywhere—#console, #reminders, #priority-feed, or #general.

🏗 Core Architecture

Dex operates on a Proposal-Execution-Verification split architecture.

  1. Conversational Ingestion: Mention @Dex anywhere or use Slash Commands.
  2. Reasoning: Gemini 2.0 Flash (or Groq) generates a structured ExecutionPlan.
  3. Safety Gate: The Risk Engine audits every step for potential system impact.
  4. Confirmation: High-risk tasks pause for a native Discord button confirmation.
  5. Autonomous Execution: The ExecutorAgent runs Python-based tools locally.
  6. Branded Verification: Results are verified and posted as rich embeds across your server.

🚀 Key Features

⚡ Hermes Messenger Layer

  • Server-Wide Interactivity: Dex responds to natural language pings in all designated channels.
  • Dynamic Branded Digest: A unique, LLM-generated Morning Intel Digest at 08:00 AM IST with logo thumbnails and professional formatting.
  • Multi-Channel Delivery: Seamless handoff between Discord and professional HTML emails.

🧠 Master Task Management

  • Reminders: Precision scheduled time-based alerts.
  • TODOTool: Persistent daily task tracking and priority management.
  • Integrated Intelligence: Your daily summary combines reminders, TODOs, and system health telemetry.

📡 System Operations

  • Local Memory: Semantic retrieval of your notes and past task history.
  • Telemetry: Real-time performance monitoring via /dex telemetry.
  • Render Stability: Automated keep-alive loops and process auto-restart for 100% uptime.

📦 Deployment & Setup

Render Deployment (Recommended)

Dex is optimized for Render, utilizing a "Double Shield" of stability for the free tier.

  1. Fork/Clone the repository.
  2. Create a new Web Service on Render (Docker Runtime).
  3. Set the following Environment Variables (see dex.env for full list):
    • RENDER_EXTERNAL_URL: Your public service URL (required for logos).
    • GEMINI_API_KEY: Google AI Studio key.
    • GROQ_API_KEY: Groq console key (backup brain).
    • DISCORD_BOT_TOKEN: From Discord Developer Portal.
    • NOTIFY_RESEND_ENABLED: true (use Resend for email delivery on Render).

Local Docker Setup

docker-compose up -d --build

✅ Implementation Roadmap

Feature Status Description
Hermes Interactivity Server-wide conversational support
Dual-Core LLM Gemini 2.0 + Groq (Llama 3.3) fallback
TODO Engine Persistent task management system
Branding Kit Professional SVG logo and banner integration
Dashboard UI React-based visual telemetry interface
Voice Activation Real-time voice command processing

🛡 Security & Ethics

  • Deterministic Guardrails: LLMs never execute code directly; they propose plans for the deterministic Executor.
  • Local-First Privacy: Your task data stays in .agentic_os/ on your host environment.
  • Human-in-the-Loop: Mandatory button clicks for all system-altering operations.

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.


Built with ❤️ for the future of personal AI.

About

A production-grade local-first AI agent system that helps you get things done through intelligent task planning, execution, and verification. Designed to listen, remember, and assist.

Resources

Security policy

Stars

0 stars

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors