DARKFLUX
DEV HUB

What Is Darkflux?

Darkflux V0.1 is a personal AI assistant I built entirely on my Android phone using Termux, Python, and a collection of powerful AI APIs. It's not just a chatbot — it's an automation engine that can control hardware, read the news, respond with a human-like voice, and execute hundreds of commands.

The Core Stack

• Gork AI — The reasoning brain powering all conversational and decision-making tasks.
• ElevenLabs — Ultra-realistic voice synthesis so Darkflux can speak responses aloud.
• GNews API — Real-time news streaming for market updates, tech news, and more.
• Termux:API — Direct access to Android hardware: flashlight, camera, vibration, TTS.

Why Mobile?

Building on mobile isn't a limitation — it's a choice. I wanted to prove that powerful AI systems don't require a beefy laptop or a server rack. Darkflux runs entirely on a mid-range Android phone, making it accessible to developers everywhere.

Setting Up Python in Termux

Termux transforms your Android phone into a Linux environment. Here's how to set it up and make your first API call.

pkg update && pkg upgrade -y
pkg install python
pip install requests

import requests

url = "https://api.example.com/data"
headers = {"Authorization": "Bearer YOUR_API_KEY"}

response = requests.get(url, headers=headers)
print(response.json())

Nepal IPO Market Overview

The Nepal Stock Exchange (NEPSE) continues to see strong IPO activity in Q1 2025. Several companies across banking, hydropower, and tech sectors have announced upcoming public offerings.

Bitcoin Trend Analysis

Bitcoin has shown remarkable resilience in 2025, consolidating above key support levels. Using Darkflux's GNews API integration, I track real-time crypto sentiment and price movements directly from my phone.

✅ Safe Daily Commands

• pkg update — Updates package lists safely
• python script.py — Runs Python scripts
• termux-battery-status — Checks battery level
• curl URL — Makes HTTP requests

⚠️ Power Commands (Use With Care)

• chmod 755 file — Changes file permissions
• termux-torch --on — Activates flashlight
• termux-vibrate -d 1000 — Triggers vibration

☠️ Danger Zone (Think Before Running)

# NEVER run these without understanding them:
rm -rf ~/          # Deletes everything in home dir
chmod -R 777 /     # Removes all security permissions
dd if=/dev/zero    # Can fill storage instantly

The 10 Secret Commands

These Termux:API hardware commands are the backbone of Darkflux V0.1's physical interaction capabilities.

# 1. Toggle Flashlight
termux-torch --on / --off

# 2. Text-to-Speech
termux-tts-speak "Darkflux Online"

# 3. Vibrate
termux-vibrate -d 500

# 4. WiFi Scan
termux-wifi-scaninfo

# 5. Battery Status
termux-battery-status

# 6. Screenshot
termux-screenshot -f /sdcard/shot.png

# 7. Notification
termux-notification --title "Darkflux" --content "Done"

# 8. Clipboard
echo "text" | termux-clipboard-set

# 9. Location
termux-location

# 10. Call Log
termux-call-log

The Neon-Tech Aesthetic

The Darkflux UI was designed with one goal: make it feel like you're operating a future-military AI system — but from a phone. The design language draws from cyberpunk aesthetics, terminal UIs, and dark-mode developer tools.

Core Design Principles

• Deep Space Palette — #050810 background creates depth without being pitch black
• Accent Colors — #00f5d4 (teal) for primary actions, #7b2fff (purple) for secondary
• Space Mono — Monospace font gives the terminal feel throughout
• Clip-Path Polygons — Angled corners replace standard rounded borders

/* Darkflux CSS Variables */
:root {
  --bg: #050810;
  --accent: #00f5d4;
  --accent2: #7b2fff;
}

Introduction

An AI is only as good as the data it can access. For Darkflux V0.1, I wanted a system that doesn't just guess — it actually communicates with a dedicated backend. This is where MCP (Model Context Protocol) comes in.

Today, I'll show you how I connected my custom chatbot to a live MCP server, giving it real-time capabilities and a persistent connection using SSE (Server-Sent Events).

1. What Is an MCP Server?

Think of an MCP server as the "Central Nervous System" of your AI. It allows the chatbot to:

• Perform actions — Execute real tasks based on user requests
• Fetch live data — Pull current prices, news, status updates in real time
• Maintain stable connections — Using SSE (Server-Sent Events) for streaming
• Handle context — Remember conversation state across sessions

2. Checking the Heartbeat (The Ping)

Before the chatbot starts talking, it needs to know if the server is alive. We use a simple ping endpoint to verify connectivity.

How it works: The chatbot sends a GET request to this endpoint. If it receives a 200 OK, the "System Online" indicator turns green. This happens automatically when you load the chatboard.

async function checkStatus() {
  try {
    const res = await fetch("https://your-mcp-server.com/ping");
    const data = await res.json();
    return data.status === "OK";
  } catch (e) {
    return false; // Server offline
  }
}

3. The Conversation Flow (SSE Integration)

Unlike normal APIs that give a one-time answer, SSE (Server-Sent Events) allows the server to stream the answer word-by-word. This makes the AI feel much faster and more natural — you see text appear as it's generated.

async function streamResponse(query) {
  const response = await fetch("/sse", {
    method: "POST",
    headers: { "Content-Type": "application/json" },
    body: JSON.stringify({ query })
  });

  const reader = response.body.getReader();
  const decoder = new TextDecoder();
  let result = "";

  while (true) {
    const { done, value } = await reader.read();
    if (done) break;
    result += decoder.decode(value);
    updateChatBubble(result); // Show partial text
  }
}

Use a standard text input with proper mobile keyboard support:

<input type="text" id="user-query"
       placeholder="Ask anything..."
       inputmode="text"
       autocomplete="off"/>

Run a local test server directly from your Android phone:

# Start local HTTP server in Termux
python -m http.server 8080

# Your server is now at:
# http://localhost:8080

4. Why This Matters for Darkflux

By using this MCP architecture, Darkflux stays lightweight on your phone while the heavy processing happens on the Render cloud. This design means:

• No phone hanging — All AI computation happens server-side
• Battery savings — Streaming is more efficient than polling
• Real-time responses — SSE delivers text as fast as it's generated
• Scalable — Add more tools to the MCP server without updating the app

5. Conclusion

Connecting your UI to an MCP server is the final step in creating a professional-grade AI assistant. It's the difference between a simple toy and a powerful tool. The chatboard below demonstrates a working AI powered by a real backend — try it yourself!