KYBERNESIS
HOW_IT_WORKS
kybernesis://overview
READING_MODE
// INTRODUCTION

How Kybernesis Works

The second brain for your AI agents that learns, organizes, and retrieves your knowledge autonomously.

DATA_FLOW:
UploadProcessStoreOrganize(sleep)RetrieveAI Agents

The Big Picture

Imagine having a brilliant assistant who reads everything you give them, remembers it perfectly, organizes it while you sleep, and can instantly recall exactly what you need when you ask.

That's Kybernesis.

01

Upload Your Knowledge

You can bring information into Kybernesis in three ways:

Direct Uploads

Drop in PDFs, Word documents, spreadsheets, or text files. We handle the messy work of extracting text, even from scanned documents using OCR.

Chat Ingestion

Paste in conversations, notes, or ideas. Great for capturing quick thoughts or importing chat logs.

Connected Apps

Link your Google Drive or Notion, and we'll automatically sync your documents. New files get pulled in on a schedule.

02

Smart Processing

When a document arrives, several things happen automatically:

TEXT_EXTRACTION

Whether it's a PDF, Word doc, or spreadsheet, we convert it into clean, searchable text. Scanned PDFs? We use AI-powered OCR to read the images.

INTELLIGENT_CHUNKING

Long documents get broken into digestible pieces—roughly 500 words each. We're careful to break at natural boundaries (sentences, paragraphs) so ideas stay together.

SEMANTIC_EMBEDDING

Each chunk gets converted into a mathematical representation (an "embedding") that captures its meaning. This lets us find information by what it's about, not just by keyword matching.

03

Dual Storage

Your memories live in two places, working together:

VECTOR_STORE

Semantic Memory

Stores the mathematical meaning of your content. This powers semantic search—finding things that are conceptually related, even if they don't share exact words.

STRUCTURED_DB

Metadata & Relationships

Stores metadata, tags, access history, and relationships. This powers filtering by date, source, or category.

Think of it as having both a librarian who understands concepts AND a filing system that tracks every detail.
04

The Sleep Agent

MAGIC

Here's where the magic happens. Every hour, while you're working (or actually sleeping), our AI agent quietly maintains your knowledge base:

TAG

Auto-Tagging

The Sleep Agent reads memories that haven't been categorized and assigns relevant tags. A document about "quarterly revenue projections" might get tagged withfinanceQ4projectionsplanning.

LINK

Relationship Building

Memories that share themes get connected. Your project proposal might link to related research notes, team meeting summaries, and budget documents—even if you never explicitly connected them.

TIER

Priority & Tiering

Not all memories are equal. Recent, frequently-accessed, highly-connected memories stay in thehot tier for instant retrieval. Older, isolated memories move to warm thenarchive tiers. Nothing is deleted—just organized by relevance.

SUM

Summarization

Long documents get condensed summaries for quick previews. You can scan what a memory is about without reading the whole thing.

05

Retrieval

When you search for something, we use a hybrid approach:

70%
Semantic Search

"What's our approach to customer onboarding?" finds documents about onboarding even if they don't use that exact phrase.

30%
Metadata Filtering

Narrow by source, date range, tags, or tier. "Show me onboarding docs from last month" combines meaning with structure.

Results come back ranked by relevance, with snippets showing why each memory matched.

06

AI Agents

This is where Kybernesis becomes truly powerful. You can create AI agents that use your memories as their knowledge base.

How Agents Work

Each agent has a personality (defined by you), access to your memories, and the ability to have natural conversations. Ask an agent a question, and they'll search your knowledge base to give informed answers.

Agent Memory

Agents remember your conversations. They track who you are, what you're working on, and your preferences. This context persists across sessions.

EXAMPLE_CONVERSATION:
YOU:"What did we decide about the pricing model last quarter?"
AGENT:"Based on your October strategy doc and the finance team's analysis, you settled on a tiered pricing model with three levels. The key decision was to include a free tier with 30 memory limit to drive adoption. Should I pull up the specific documents?"

The Topology View

One unique feature is the Topology—a visual map of your memories and their connections.

Nodes represent memories. Edges represent relationships. Clusters form naturally around themes. You can see how your knowledge interconnects, discover unexpected links, and manually add connections you find meaningful.

It's not just storage—it's a living knowledge graph that evolves over time.

Privacy & Security

Your memories are yours.

  • All data is encrypted in transit and at rest
  • Multi-tenant isolation ensures your data is completely separate from other users
  • OAuth connectors only access files you explicitly authorize
  • API keys let you control programmatic access

What Makes This Different

It's not just search.

Traditional search finds documents with matching words. Kybernesis understands meaning and context.

It learns while you sleep.

The Sleep Agent continuously improves organization without any effort from you.

It builds connections.

Memories don't exist in isolation—they form a knowledge graph that reveals insights.

It powers AI agents.

Your knowledge becomes the foundation for AI assistants that actually know your context.

Getting Started

  1. 1
    Uploada few documents to get started
  2. 2
    Watchthe Sleep Agent organize them (check back in an hour)
  3. 3
    Searchusing natural language queries
  4. 4
    Createan agent to chat with your knowledge base
  5. 5
    ConnectGoogle Drive or Notion for automatic syncing

Your second brain is ready when you are.

Kybernesis: Remember everything. Find anything. Understand the connections.