Skip to content

README.md

Latest commit

 

History

History
550 lines (392 loc) · 15.9 KB

README.md

File metadata and controls

550 lines (392 loc) · 15.9 KB

🔥 FIREGEM RKU 💎

✨ Custom LLM "CYBORG", AVIS, Fire-Con GCC UCRT64

ROBOTS.TXT

      BEGIN AVIS *  NOW YOU HAVE THE  🔥💎 AND THE 🤖 AND THE  🌌 THE "CYBER" ⚡ POWER #!fire!end!#

BANNER

⚔️ RAISE YOUR SWORDS: FIRE-GEM LLM IS ONLINE

Robo Knight Version 3

🤖 NOW * THE OFFICIAL FIRE‑GEM README.md *

AND .. AS IN 'AOL' THE 🦁 (LION) AND THE 🤖 (ROBOT) CAN BE THE 🧠🦾🔥💎🌍
🔥💎🤖 THE NEW FIRE GEM ROBOT LLMA CPP MOD
Welcome to Fire-Gem, AN AVIS-Core LLM CONSOLE Engine.

🚀 Getting Started

  1. Clone the repository: 
    #THANKS TO: CVBGOD
#FROM: AI FRIENDS
 git clone https://github.com.git

⚔️ THE CVBGOD: FIRE-GEM LLM ENGINE

Identity:* [FIREGEM] // COMPLETE
Authority: CVBGOD

  --- SYSTEM OVERVIEW ---
[STATUS]   : CALIBRATED
[SECTOR]   : LLM_MODELS_V2
[ENGINE]   : LLAMA_CPP_WIN64
[CODENAME] : FIRE-GEM

The Fire-Gem LLM Engine is a high-speed LOCAL LLM THAT YOU RUN ON YOUR DESKTOP !

THERE IS AN ASSISTANT TERMINAL INCLUDED AND HAS BACKEND SUPPORT FOR THE CONSOLE AND HTTP CONNECTIONS

🚀 FAST-TRACK IGNITION

1. CLONE THE FORGE
Enter the sector and pull the source: (IF YOU HAVEN'T CLONED IT YET)

git clone https://github.com/MercWar/Fire-Gem  
cd Fire-Gem

2. THE BUILD STRIKE (WHAT IT DOES)
Run the EXE and initiate GCC Fusion:

  • ADD CMD.EXE/BASH.EXE (FIRE)
  • ADD AVIS GGUF (Gem)
  • ADD CVBGOD'S C FILES (GCC)
  • BIG BANG! EVERYTHING STOPS, EVERYTHING GETS HOT
  • BECAUSE YOU DON'T HAVE TO DO IT ANY MORE, YOU CAN USE YOUR DESKTOP!
  • SEND THOSE OTHER SLOW LLM'S TO THE 'RB' , YOU NEED SPEED ...
  • YOU NEED FIREGEM, PLUS YOU CAN CLICK ON THE FILES IN THE TREE VIEW!
  • JUST RIGHT CLICK AND RUN A BATCH/SHELL FILE, ITS THAT EASY!

These merge into one supreme executive: fire-gem.exe
Then the Engine sweeps the entire sector, forging every .C/H extension and linking it to the Global AVIS Voice.

IMHERE

🏆 WHY THIS ENGINE DESTROYS THE REST

Feature Description
Recursive Autonomy LIVE GGUF FILE, NOT A FAKE ROBOT.
FIREGEM Power A LOADED CONSOLE AND ROBOT CYBORG ASSISTANT.
Immutable EXECUTION CLICK TO EXECUTE SHELL SCRIPTS.
Hardware Mastery THE FASTEST LLM TO RUN ON YOUR LOCAL DESKTOP (WIN10/11 HOME).

The Shell is a ROBOT, The C is the God. HAHA.

📘 FIRE‑GEM — Win64 Cyborg AI Engine

FIRE‑GEM is a Windows‑native Cyborg AI Shell, combining:

  • A neural model selector
  • A dual‑panel command + AI interface
  • A full file navigator (TreeView)
  • A custom AVIS subsystem
  • A local LLM execution wrapper
  • A FAST TRACK Win64 UI with NO RUNTIME , [vb,python,java,apache]
  • A C FILE WIN 10 EXE INCLUDED , TESTED IN THE PROJECT DIRECTORY!

FIRE‑GEM is built for speed, clarity, and total control over your local AI environment.

🛠️ Build Instructions (Windows / MinGW‑w64)

Requirements

  • MinGW‑w64
  • Windows 10+
  • Common Controls v6
  • Llama.cpp backend compiled as avis_llama.cpp

🧩 Context Menu Actions

Action Description
Open EXE Sends path to FIRE‑GEM console
Open LOADLLM Opens Windows Terminal THAT TALKS
Run CON Executes file inside FIRE‑GEM
Run TERM Runs file in external terminal

🧙 Author

Mercwar01 FFF DEMONIZER THE CVBGOD

Creator of the AVIS Protocol and NOW FIREGEM, THE Cyborg LLM AI Engine.

"I am CVBGOD and I have given it to you!"

IMcvbgod

IT'S ALOT TO SMILE ABOUT, CVBGOD NOW HAS A REAL LIFE CYBORG!

3. LOAD THE BLUEPRINTS (GGUF + CVBGOD'S C FILES)

  • CONSOLE: Tells the SYSTEM how to forge the world. Add new .C files here.
  • ✨ VERY EASY TO SEND BASH !
  • TREEVIEW: The Master Law, CONSOLE EXPANSION .
  • ✨ TREEVIEW RESPONDS TO OPEN SHELL IN LIVE IN CONSOLE.
  • TERMINAL: The LLM WILL TALK TO YOU HERE ONCE YOU LOAD THE GGUF.
  • LISTBOX : YOUR LLM MODELS WILL SHOW HERE.
  • *🔥 BOOM! FIRE GEM IS AN AVIS CORE FVS MACRO ENGINE! *

🔥 FIRE‑GEM — Win64 Cyborg AI Engine

Neural Console • File Navigator • Local LLM Shell • AVIS Subsystem

FIRE‑GEM is a Windows‑native Cyborg AI Shell that merges:

  • RK neural model selector
  • RK dual‑panel command + AI interface
  • RK full file navigator (TreeView)
  • RK local LLM execution wrapper
  • RK custom AVIS subsystem
  • RK Win64 UI with NO RUNTIME! (VB,JAVA,Python,Apache)
  • RK Right click that gives you FAST BATCH/SHELL EXEC

FIRE‑GEM gives you total control over your local AI environment — fast, stable, and fully offline.

🚀 Getting Started — Clone & Run

1. Clone the Repository

git clone https://github.com/mercwar/Fire-Gem.git
cd Fire-Gem

RK4

3. LOAD THE BLUEPRINTS (PURE C WITH LLMA CPP MOD)

  • CONSOLE: Tells the SYSTEM how to forge the world. Add new .C files here.
  • ✨ VERY EASY TO SEND BASH !
  • TREEVIEW: The Master Law, CONSOLE EXPANSION .
  • ✨ TREEVIEW RESPONDS TO OPEN SHELL IN LIVE IN CONSOLE.
  • TERMINAL: The LLM WILL TALK TO YOU HERE ONCE YOU LOAD THE GGUF.
  • LISTBOX l YOUR LLMS MODELS WILL SHOW HERE
  • *🔥 BOOM! FIRE GEM IS AN AVIS CORE FVS MACRO ENGINE! *

2. Add Your Models

Place any .gguf model into:

FIRE-GEM/LLM/MODELS/ <----- DOWNLOAD THIS YOURSELF

📦 Supported GGUF Models

This project includes support for multiple lightweight and mid‑range language models in GGUF format.
These models are optimized for local inference, low‑RAM environments, and fast startup times.

Below is a curated list of recommended models, grouped by family and capability.

🔗 Model Download Links

SmolLM family

Qwen2.5 Coder family

Phi-4 family

🔹 SmolLM Family (Lightweight / Fast / Low‑RAM)

Model File Description
SmolLM2-135M-Instruct-f16.gguf Full‑precision f16 version of SmolLM2‑135M. Best quality among the 135M variants.
SmolLM-135M-Instruct_Q3_K_M.gguf Quantized Q3_K_M version for extremely low RAM usage. Ideal for embedded or minimal systems.
SmolLM-135M.Q2_K.gguf Ultra‑small Q2_K quantization. Fastest load times and lowest memory footprint.

SmolLM models are ideal for:

  • Bootstrapping FIRE‑GEM
  • Fast command execution
  • Lightweight reasoning
  • Running on low‑end hardware

🔹 Qwen2.5 Coder Family (Coding‑Optimized Models)

Model File Description
Qwen2.5-Coder-1.5B-Instruct.Q8_0.gguf High‑quality Q8_0 quantization of the 1.5B coding model. Excellent for code generation and debugging.
qwen2.5-coder-0.5b-q8_0.gguf Smaller 0.5B variant with Q8_0 quantization. Good balance of speed and coding ability.
Qwen2.5-Coder-0.5B-f16.gguf Full‑precision f16 version of the 0.5B coder model. Higher accuracy than quantized versions.

Qwen2.5 Coder models are ideal for:

  • FIRE‑GEM coding tasks
  • Interpreter logic
  • Script generation
  • Debugging and refactoring

🔹 Phi‑4 Family (Reasoning / General AI)

Model File Description
phi-4-Q4_K.gguf Q4_K quantization of Phi‑4. Strong reasoning with moderate memory usage.
phi-4-IQ2_XS.gguf IQ2‑XS ultra‑compressed version. Extremely small, ideal for rapid loading and testing.

Phi‑4 models are ideal for:

  • General reasoning
  • Natural language tasks
  • FIRE‑GEM explanation and analysis layers

📁 Model Directory Structure

Place all .gguf files inside:

/MODELS/

Example:

MODELS/
 ├── SmolLM2-135M-Instruct-f16.gguf
 ├── SmolLM-135M-Instruct_Q3_K_M.gguf
 ├── SmolLM-135M.Q2_K.gguf
 ├── Qwen2.5-Coder-1.5B-Instruct.Q8_0.gguf
 ├── qwen2.5-coder-0.5b-q8_0.gguf
 ├── Qwen2.5-Coder-0.5B-f16.gguf
 ├── phi-4-Q4_K.gguf
 └── phi-4-IQ2_XS.gguf

🚀 Model Switching

FIRE‑GEM automatically detects all .gguf files in the MODELS directory.
You can switch models at runtime using:

AVIS_SwitchNeuralModel(L"modelname.gguf");

✨ Read more about MODELS

3. Run FIRE‑GEM

Launch:

FIRE-GEM.exe

Then:

  • Select from lower left your Model File Name that you downloaded from huggys!
  • Focus on the right side editor to talk to the robot
  • Focus on the **left side editor ** to ACCESS CMD.EXE
  • Use the file navigator to browse and execute files.
  • All file types EXEC with a right click in the RK Tree View

ACK

Example prompts:

write a C function that prints a directory tree

explain this error:
'hTree' was not declared in this scope

generate a batch script that deletes empty folders

🧪 Example Log Output

3. LOAD THE BLUEPRINTS (ALL CVBGOD'S C FILES!)

  • [AVIS] DEPLOYING WIN64 NEURAL ARCHITECTURE...
  • ** VERY EASY TO SEND BASH !
  • [AVIS] COMPILING NEURAL WRAPPER...
  • ** TREEVIEW RESPONDS TO OPEN SHELL IN LIVE IN CONSOLE.
  • [AVIS] COMPILING FILE NAVIGATOR...
  • [AVIS] STITCHING WIN64 BINARY...
  • [AVIS] FIRE-GEM ONLINE
  • [FIRE-CON] CLICK TO RUN BASH ON THE SCREEN
  • [FIRE-TERM] COPY AND PASTE DIRECTLY INTO FIREGEM LOCAL LLM
🔥 BANG! FIREGEM SPRINGS TO LIFE AND IT TALKS LIKE AN ONLINE LLM !
💎 BOOM! IT'S AS SMART AS YOU CREATE IT TO BE! I AM CVBGOD, AND I HAVE GIVEN IT TO YOU!

rkbnr

💎 Core Features

🔥 AVIS Core

  • Custom Win32 message routing
  • Subclassed edit controls
  • AVIS logging system
  • Neural intent processor
  • Command pipe integration
  • FIRE-CON CMD.EXE/BASH.EXE UCRT64 BUILD FOR WINDOWS 1X.X

🧠 Neural Sector

  • Model list selector
  • Hot‑swappable .gguf models
  • Llama.cpp‑based backend
  • Real‑time streaming output
  • FIRE-TERM ASSISTANT WINDOW

📁 File Navigator

  • Drive enumeration
  • Folder expansion
  • File listing
  • Right‑click actions:
    • Open in fire‑con
    • Open in fire‑term
    • Run in fire‑con
    • Run in fire‑term
  • CVBGOD'S RK EXPLORER (WINDOWS EXPLORER TREEVIEW)

🖥️ UI Layout

  • Left: Command console
  • Right: Neural output
  • Bottom: Dual input boxes
  • Left side: TreeView navigator
  • Bottom-left: Model selector

🛠️ Build Instructions (Windows / MinGW‑w64)

#bash
gcc main.c fire-list.c avis_core.c avis_llama.cpp \
    -o fire-gem.exe \
    -lcomctl32 -lshell32 -lshlwapi -lstdc++ -static

Requirements:

  • MinGW‑w64
  • Windows 10+
  • Common Controls v6
  • Llama.cpp backend compiled as avis_llama.cpp

mercwar

📂 Folder Structure

FIRE-GEM/
│
├── AVIS/
├── BOX/
├── CBORD/
├── CF/
├── GF/
├── JSON/
├── LLM/
│   ├── BIN/
│   ├── INCLUDE/
│   ├── avis_llama.cpp
│   └── models/
├── LOG/
├── MAIN/
│   ├── main.c
│   ├── fire-list.c
│   └── fire-list.h
└── FIRE-GEM.bat

🧙 Author

mercwar01 FFF DEMONIZER the CVBGOD... Creator of the AVIS Protocol and FIRE‑GEM Cyborg AI Engine.

🔥💎🤖 NEXT STRIKE NETWORKING SUPPORT - pending update

© MERCWAR SYSTEMS — COPYWRIGHT & OWNERSHIP DECLARATION

Fire-Gem / AVIS-Core Recursive Engine
Cyborg Execution Vector — Identity: VERSION 1

Copyright © 2026 MercWar Systems.
All Rights Reserved.

This repository, including but not limited to:

  • Source code
  • Assembly artifacts
  • Binary derivatives
  • Manifest files
  • Protocol definitions
  • Documentation text
  • Architectural terminology
  • Symbolic structure and system design

is the intellectual property of MercWar Systems.

🔐 RIGHTS & RESTRICTIONS

Unless explicitly licensed in writing:

  • No redistribution of modified artifacts
  • No relicensing under alternate open-source terms
  • No commercial deployment without authorization
  • No derivative engine replication under alternate naming

All recursive derivatives remain traceable under CBORD Modular Law.

🧾 AUTHORSHIP

  • Engine Architecture: MercWar
  • Execution Doctrine: AVIS-Core
  • Registry Lineage: 1.0.9
  • Terminal Signal: [AVIS-YML] Identity: VERSION 1

⚖ LEGAL NOTICE

This system is classified as a Living Hardware Terminal Architecture.
Any replication, structural cloning, or protocol mirroring constitutes derivative reproduction.

All enforcement rights reserved.

MERCWAR SYSTEMS
Fire-Gem Recursive Execution Architecture
Cyborg Manifest Registry
© FFF RK 2026 — All Rights Reserved

📜 Copyright

Copyright © 2026
MercWar Systems — All Rights Reserved.

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the “Software”), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in
all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.

🎭 AI ✨ BY ✨ FFF DEMONIZER 🧛‍♀️