Documentation
Everything you need to ship.
The complete index — getting started, API reference, MCP integration, security, and the architecture story behind it.
Start here
Getting started
From sign-up to your first AI bundle in fifteen minutes. Read in order, or jump straight to the part that matches what you're building.
Quickstart
Sign in, make a box, write a note, and bundle it for an AI in five steps.
Concepts: boxes, notes, links
The smallest mental model that unlocks the rest of the product.
Concepts: skills & agents
Real package structure for the work you do with AI — not flat blobs.
Importing from Obsidian / Notion
Bring your existing knowledge in. Export back out at any time.
Build
API, MCP, and integrations
Everything you need to script Poggle, ship MCP tools, and wire it into the rest of your stack.
Send context to your AI
Bring your Poggle notes into Claude, Cursor, ChatGPT — any AI tool — in one paste.
REST API v1 reference
OAuth 2.1 authentication, scoped tokens, rate limits, and every endpoint.
Model Context Protocol (MCP)
First-class MCP server. Plug Poggle into Claude, Cursor, or any MCP-aware client.
Connections & permissions
Per-connection scopes, write-with-approval mode, audit log on every call.
Branch-aware writes
How AI writes land on a branch first, with diff review and optional CI gates.
Operate
Trust, security, and governance
What we built so this product can be trusted with the work — and the controls available to your team.
Trust & security overview
Eight pillars: data isolation, auth, keys, audit, branches, encryption, network, portability.
Privacy policy
What we collect, what we don't, and how to ask for everything back.
Acceptable use
The short list of things you cannot do with Poggle.
Organization & team controls
Members, roles, retention, and enterprise SSO / SCIM.
Reference
Deep dives
When you want the full architecture story — for a security review, a build-vs-buy decision, or just curiosity.
Architecture overview
Hex/onion separation, server components, RLS, MCP, workers, and CRDT.
Knowledge graph
Entities, semantic links, and the GraphRAG context the operator agent uses.
Agents & operators
How long-running agents run on a branch, propose writes, and report back.
Object lifecycle
Draft → active → archived → trashed, plus version history on every object.
Stuck? We're a fast email away.
Most replies inside one business day. Security questions go to a dedicated address with the SIG-Lite, DPA, and architecture diagrams pre-loaded.