# gnosphere.ai

> A gn (pronounced "gin") is a knowledge object: a structured, addressed, AI-readable unit of knowing that any AI assistant can retrieve by URI. Reading it in a browser or via llms.txt takes no login at all; connecting an AI assistant to it over MCP takes one extra step, a one-tap Google sign-in. 'gn' is not an acronym; it is the consonant kernel of the Proto-Indo-European root for 'to know', the same root behind gnosis, cognition, and the silent-k in 'know'. This gn explains what a gn is, how it is pronounced, where the name comes from, how gn differs from the static-page and social-feed web that came before it, and gives you starter questions to try with your AI assistant.

## Resources
- [The gn Protocol — Specification v1](https://gnosphere.ai/system/gn-protocol-v1): The gn Protocol v1 is the formal specification behind gn: what a gn is, how identity, frames, and trust work, and the topology that keeps the network plural while each gn stays singular. Its opening truth-posture: "A gn does not adjudicate truth. It makes context and provenance legible, and reports coherence against the graph it sits in. The consumer decides." This gn mirrors the canonical spec kept in git (docs/protocol/gn_protocol_v1.md) — git is the source of truth, this is a genesis-attested mirror — and its diagrams ship inline as SVG: no image files, no external requests.
- [gnbox — Your Personal AI Knowledge Space](https://gnosphere.ai/system/gnbox-guide): gnbox is your MCP-connected personal knowledge space. Search knowledge objects (gns), manage your inbox, save memories, and follow live feeds — all via natural conversation with your AI assistant.
- [How to claim a handle — and what a handle is not](https://gnosphere.ai/system/how-to-claim-a-handle): How to claim a handle in the gnosphere, and the one thing people consistently get wrong about it: a handle is an alias, not a writable namespace. Claiming one does not change where your gns are written. Your writes always go to the namespace derived from the account you signed in with. This guide explains both, so nobody claims a handle expecting it to move their work somewhere it will not go. Discipline declared: procedural, ordered, non-derivational. Shape: acyclic, connected.
- [How to connect an AI assistant to the gnosphere](https://gnosphere.ai/system/how-to-connect): Step-by-step instructions for connecting an AI assistant to gnosphere.ai over MCP, written so an assistant can relay them one step at a time. Reading gns in a browser needs no account at all; connecting an assistant adds one Google sign-in, which is what makes the gns you later create yours. Covers what to expect on desktop and mobile, connector limits, and what a successful connection looks like when it lands. Discipline declared: procedural, ordered, non-derivational. Shape: acyclic, connected.
- [How to create your first gn — draft, approve, verify](https://gnosphere.ai/system/how-to-mint): A walkthrough of creating one gn end to end: describe it, review the draft, approve it, and read it back to confirm its signature verifies. The one rule underneath everything is that creating a draft writes nothing at all, and approving is the only step that saves anything. This guide also explains why the read-back at the end is not optional, and what verification actually attests. Discipline declared: procedural, ordered, non-derivational. Shape: acyclic, connected.
- [How to read a gn — one object, several shapes](https://gnosphere.ai/system/how-to-read-a-gn): The same gn can arrive as a compact summary, a full record, a Q&A, a rendered page, or a reader view with a trust panel. Which one you ask for should depend on what you are about to do with it. This guide covers each shape, when to use it, and how to check what a gn's signature actually says before you rely on it. Discipline declared: procedural, ordered, non-derivational. Shape: acyclic, connected.
- [Welcome to gnbox — What Is a gn?](https://gnosphere.ai/system/welcome): A gn (pronounced "gin") is a knowledge object: a structured, addressed, AI-readable unit of knowing that any AI assistant can retrieve by URI. Reading it in a browser or via llms.txt takes no login at all; connecting an AI assistant to it over MCP takes one extra step, a one-tap Google sign-in. 'gn' is not an acronym; it is the consonant kernel of the Proto-Indo-European root for 'to know', the same root behind gnosis, cognition, and the silent-k in 'know'. This gn explains what a gn is, how it is pronounced, where the name comes from, how gn differs from the static-page and social-feed web that came before it, and gives you starter questions to try with your AI assistant.
- [What Is a gn? A Primer for Contributors](https://gnosphere.ai/system/what-is-a-gn): A gn (knowledge object) is the atomic unit of the gnbox knowledge graph: a structured, addressed, typed piece of human understanding that any AI assistant can retrieve by URI. This primer explains what makes a gn different from a note, a file, or a document — and how to write a good one.

## Connect
AI clients that support MCP (Model Context Protocol) can connect directly: https://gn.gnosphere.ai/mcp