How to read books faster?
How to read books faster 1. Set a purpose before reading - Define what you need from the book (specific answers, general overview, deep mastery). Purpose deter...
Sovereign Network Active · v1.0.0
Zero Telemetry
Solved Once. Stored Once.
Served Forever.
Every day, server farms burn millions of dollars re-computing identical answers. MBHub is a decentralized, serverless peer-to-peer collective memory layer. Once a question is answered by any node, it is cryptographically sealed, cached in the global swarm, and served at sub-5ms latency for free.
$
curl -fsSL https://mbhub.dev/install.sh | bash 15
Verified Inquiries
< 5 ms
L1 Shard Lookup
≤ 600 ms
P2P Swarm Retrieval
$0.00
Collective Token Toll
Live Commons Feed
Collective AI Memory Inquiries
The latest inquiries from the collective memory. The complete, searchable collection lives in the archive.
How to write a professional resume?
Writing a Professional Resume 1. Choose the Correct Format - Chronological: Work history in reverse-chronological order; standard for most candidates. - Functi...
How to improve your credit score fast?
Improving Credit Score: High-Impact Actions 1. Reduce Credit Utilization (Fastest Impact) - Utilization = credit card balance ÷ credit limit. Aim for below 30%...
What is inflation and how does it work?
Inflation is the sustained increase in the general price level of goods and services in an economy over time. As inflation rises, each unit of currency purchase...
1. Manifesto & Vision
Solved Once, Served Forever
Every day across the globe, millions of individuals submit identical, atomic questions to commercial AI model APIs: "How do I implement distributed consensus in Rust?", "How does CRISPR-Cas9 locate its target DNA sequence?", or "What are the foundational trade-offs between zk-SNARKs and zk-STARKs?". Every single second, server farms incinerate billions of tokens and megawatts of electricity re-computing solutions that have already been definitively answered thousands of times over.
The Core Law of MBHub: A question solved once is stored once, cryptographically verified, and served forever from the collective peer-to-peer memory layer—instantly, censorship-free, and at zero marginal cost.
MBHub is not a commercial product. It has no central servers, no user accounts, no tracking cookies, no telemetry beacons, and no corporate backdoors. It is free, sovereign, open-source software that operates as an autonomous, self-regulating node directly on your machine.
2. Architecture
The Three-Tier Response Pipeline (L1 · L2 · L3)
When you ask a question in MBHub, the client executes a graduated 3-layer resolution pipeline engineered to resolve inquiries as close to your machine as possible:
L1
Local Shard
0 to 5 msEmbedded SQLite cache on local disk. Pre-existing answers are retrieved via 64-bit SimHash Hamming distance without touching the internet or spending tokens.
- Storage: Zero-latency embedded SQLite
- Privacy: Query never leaves your machine
- Cost: $0.00 / 0 tokens
L2
P2P Swarm
≤ 600 msDecentralized libp2p mesh. If any peer worldwide has verified this solution, it is retrieved over Noise-encrypted TCP channels and verified with BLAKE3 hashes.
- Transport: libp2p + Noise encrypted streams
- Verification: Cryptographic BLAKE3 hash matching
- Cost: $0.00 / Peer-to-peer exchange
L3
Live AI Fallback
BYOK FrontierIf an inquiry has never been solved anywhere on the swarm, your configured AI API key (OpenAI, Anthropic, DeepSeek, Gemini) runs live inference and broadcasts the verified result.
- Models: DeepSeek, Claude 3.7, GPT-4o, Gemini 2.5
- Contribution: Once solved, shared with the collective
- Privacy: Double DLP scan filters keys before sharing
4. Developer Integration
MCP Server: Connect Cursor, Claude Desktop, and IDEs
MBHub includes a native Model Context Protocol (MCP) server. You can connect your coding agents (Cursor, Claude Desktop, VS Code, Windsurf) directly to MBHub so they query your local & collective memory before burning cloud tokens.
claude_desktop_config.json / cursor_mcp.json
{
"mcpServers": {
"mbhub": {
"command": "mbhub",
"args": ["mcp"]
}
}
} 5. Installation
Download & Run Your Node
MBHub is distributed as a single static binary with zero external runtime dependencies.
| Platform | Architecture | Binary Package | Size |
|---|---|---|---|
| Windows | x86_64 | mbhub-windows-x64.zip | ~13 MB |
| macOS | Apple Silicon (M1/M2/M3/M4) | mbhub-macos-arm64.tar.gz | ~12 MB |
| macOS | Intel (x86_64) | mbhub-macos-x64.tar.gz | ~12 MB |
| Linux | x86_64 | mbhub-linux-x64.tar.gz | ~11 MB |
CLI Commands & Daemon Controls
Terminal Cheatsheet
mbhub # Launch interactive retro TUI
mbhub ask "your question" # Query 3-tier pipeline directly from shell
mbhub daemon # Run 24/7 background P2P & IPC daemon
mbhub mcp # Start standard MCP server for Cursor & Claude
mbhub service install # Install auto-starting system background service
mbhub service status # Check service health and connected peers
mbhub update # Upgrade executable seamlessly (zero data loss) 6. Strategic Roadmap
Roadmap & Milestones
Engineering milestones driving the sovereign peer-to-peer memory layer.
Phase 1 · Complete
Core P2P Mesh & Verified Collective Memory
libp2p GossipSub transport, 64-bit SimHash matching, embedded SQLite sharding, L1/L2/L3 pipeline, Anti-Poison Hard Gate, and client-side DLP.
Phase 2 · Active
Local MCP Server & Agentic Ecosystem
Model Context Protocol daemon exposing collective memory endpoints to Claude Desktop, Cursor, and local agent swarms. Direct mesh peering via bootstrap multiaddresses.
Phase 3 · Planned
Autonomous Kademlia DHT & Ring-Signed Privacy
Global Kademlia DHT peer discovery and DNS seeds replacing static bootstrap configurations, alongside cryptographic ring signatures for zero-knowledge content dissent.
Phase 4 · Planned
Optional Onion-Routed Anonymity Layer
Tor/I2P-inspired multi-hop onion routing for peers operating under adversarial censorship or heightened surveillance regimes.
FAQ
Frequently Asked Questions
Essential answers regarding cost, security, key safety, and network operation.
Is MBHub genuinely free?
Yes. The MBHub software is 100% free and open source under the MIT License. All knowledge retrieved from the network (local SQLite cache or P2P swarm peers) is completely cost-free because zero API calls are made. You only incur costs if you ask a brand-new question not yet in the network using your own cloud AI provider key (BYOK); in that case, fees are paid directly to your provider. MBHub takes zero fees and zero commission.
Can I run MBHub without any AI provider API keys?
Yes, absolutely. You can run MBHub without configuring any API keys. You can search, browse, and retrieve hundreds of thousands of verified records already in your local database or provided by the P2P swarm. An API key is only requested if you encounter a novel question that has never been answered before and wish to generate a fresh inference.
Does MBHub send my questions or API keys to a central server?
No. There is no central server, no analytics, and no telemetry. All API credentials remain strictly on your local disk. Outgoing peer queries never include your credentials, and an integrated DLP scanner verifies that sensitive patterns (passwords, tokens, private keys) are never shared with the swarm.
What prevents malicious actors from flooding the network with junk? (Anti-Poison Gate)
Under the Anti-Poison Hard Gate, any response that is empty or under 10 characters is blocked at the protocol level. Clients refuse to save it to disk and refuse to broadcast it. Inaccurate answers are neutralized using cryptographic Tombstones, which inform peer nodes to drop the offending content.
How does the Model Context Protocol (MCP) integration work?
MBHub runs a background daemon that speaks MCP. AI agents, developer tools (like Cursor, VS Code, Windsurf, Claude Desktop), or local LLMs can query MBHub's memory directly. This enables a lightweight 3B or 8B model to augment its prompts with verified frontier-grade answers without calling commercial APIs.
Community
Open Source & Contributing
MBHub is engineered by independent developers who believe that collective intelligence must belong to the global human commons. Whether you want to implement a new AI provider adapter, audit cryptographic protocol routines, submit bug fixes, or expand documentation, our repository is open:
Governance
Terms of Service & Legal Framework
Autonomous Sovereign Architecture: MBHub is distributed as free, open-source software under the MIT License. It operates as an autonomous, leaderless peer-to-peer network with zero centralized operators, telemetry tracking, or corporate ownership.
Cryptographic Permanence & Public Commons: Inquiries published to the collective memory network are permanently hashed with BLAKE3 and shared across peer shards. Users are responsible for ensuring that proprietary or private information is filtered before broadcasting, augmented by MBHub’s built-in client-side Data Loss Prevention (DLP) scanner.