Distributed P2P Net | GPL-3.0

Resucita PCs.
Conecta. Crea.

GopherOS transforms obsolete hardware into sovereign nodes of a distributed Kademlia network. Contribute local resources, host light Gopher spaces, or lease Linux power for decentralized AI.

gopheros@node:~$
🚀 Join the Hive 📋 View Roadmap
// The Crisis

The Digital Waste Epidemic

Millions of perfectly functional computers are discarded because bloated modern software demands localized resources they cannot provide.

// E-WASTE

62M tons/year

Electronic waste generated globally. Only 22% is properly recycled. The rest ends up in landfills in Ghana, India, and China.

// OBSOLESCENCE

1B PCs affected

Windows 10/11 support ending. Core 2 Duo, Core 2 Quad, Pentium D — all declared "obsolete" despite working perfectly.

// COMPUTATIONAL LOCK

Centralized AI

Modern AI tools lock your data inside corporate silos. Your private logs and terminal text train foreign, closed models.

// NETWORK BLOAT

30M+ lines of code

Linux kernel size. Built for massive servers and hyper-threaded setups. Overkill for a local node sitting in your closet.

*** GOPHEROS HIVE ACTIVE ***

Local IP: 192.168.1.45

Gopher URL: gopher://192.168.1.45/

Kad-P2P: Discovery Engine Operational

Inference Token Ledger: Connected

Active Peers: 1,402 nodes found

Uptime: 24 days, 11 hours

gopheros@swarm:~$ _

// The Solution

From E-Waste to Human Infrastructure

GopherOS isn't a stripped-down Linux distribution. It's a custom architecture tailored to turn rescue hardware into collaborative, sovereign infrastructure nodes.

  • Serve Content: Fast, distraction-free Gopher protocol hosting natively on port 70.
  • Kademlia P2P Swarm: Automatic node discovery inspired by classic decentralized networks.
  • Cooperative AI Inference: Fragment large LLMs across local layers or execute on remote servers.
  • Inference Token Economy: Earn credits by offering CPU cycles or spend them to invoke larger models.
  • Native Linux Daemon: High-performance Linux machines can join the hive to sell raw computational power.

"De basura electrónica a infraestructura humana"

// Architecture

Connector + Plug Design

Traditional kernels duplicate code for every device driver. GopherOS decouples the universal functional connector from hardware-specific data registers. 

┌─────────────────────────────────────────────────────────────┐
│  APPLICATIONS & DISTRIBUTED SWARM                           │
│  gopherd · gopherosd (Linux client) · kad-p2p · inference    │
├─────────────────────────────────────────────────────────────┤
│  INFERENCE TOKEN LEDGER (P2P Credit Balancing)               │
├─────────────────────────────────────────────────────────────┤
│  SYSCALL INTERFACE  (15 calls, stable ABI)                 │
│  open · read · write · socket · send · recv · malloc · fork │
├─────────────────────────────────────────────────────────────┤
│  KERNEL CORE  (~5,000 lines, formally verified)              │
│  scheduler · memory · paging · interrupts                   │
├─────────────────────────────────────────────────────────────┤
│  CONNECTORS  (universal, written ONCE)                     │
│  NIC Connector · Video Connector · Storage Connector       │
├─────────────────────────────────────────────────────────────┤
│  PLUGS  (hardware-specific, ~50 lines each)              │
│  RTL8139 · NE2000 · e1000 · VGA · VESA · PS/2 · IDE           │
├─────────────────────────────────────────────────────────────┤
│  HAL  (x86 specific: inb/outb, mmio, CPUID, MSRs)          │
└─────────────────────────────────────────────────────────────┘

Total kernel: ~10,000 lines  vs  Linux: 30,000,000+ lines
// Roadmap

12 Months to v1.0

We need kernel hackers, systems engineers, network researchers, and enthusiasts.

MONTH 0
Project Launch
Architecture design, repository layout, open community channels.
MONTH 1
Foundations & Assembly
Boot sector execution, 32-bit protected mode translation, basic VGA text memory buffer.
MONTHS 2-3
Functional Core
GDT/IDT mappings, paging routines, memory allocators, PS/2 peripheral drivers, and BSOD management.
MONTHS 4-5
Networking & Gopher Swarm
lwIP embedded stack integration, base NIC connectors, Gopher server listening, Kademlia DHT scaffolding.
MONTHS 6-8
Cooperative Inference & Token Engine
P2P layer fragmentation protocols, local Q8 inference tests, credit verification ledger, gopherosd Linux daemon release.
MONTHS 9-12
Live CD & Distributed Rollout
Live ISO builds, auto-hardware detection maps, unified P2P routing table initialization, stable v1.0 milestone.
// Target Hardware Matrix

Find Your Node Class

Whether you own a dead laptop or a high-end server farm, there is a specialized tier for your machine.

Hardware / Environment Cores RAM Min Role in the Swarm Economic Tier
Pentium 4 / Atom N270 1 512 MB Gopher client / Text node only Consumer Node
Pentium D / Core Duo 2 1 GB Full Gopher host + Local 20M AI Sovereign Node
Core 2 Duo / Core 2 Quad 2-4 2 GB Distributed layer processing / DHT relay Swarm Worker
Intel Core i3 / i5 (Early Gens) 2-4 4 GB Cooperative LLM Hub / Multi-layer worker Layer Master
Dedicated Linux Server / VPS 4+ 8 GB+ Heavy inference provider (`gopherosd` daemon) Token Earner ($GOPH)

Your Old PC is Not Dead.
It is Waiting.

Join a collective building alternative digital spaces — one rescued machine, one network packet at a time.

⭐ Star on GitHub 💬 Join Discord