Documents are dead.
Long live context.
We're not building another document tool. We're building the interface between human thinking and AI execution.
The Problem
For decades, we've been trained to create documents: Word files, PowerPoint decks, Notion pages. These are designed for human consumption—pretty layouts, readable fonts, carefully structured pages.
But AI doesn't read like humans. When you paste a document into ChatGPT, the AI has to reverse-engineer your intent from formatting cues. Headers become just bigger text. Layouts become noise. All that careful formatting? Mostly useless.
The real question is: what if we designed from scratch for AI consumption?
Core Principles
These principles guide every decision we make.
Context, Not Documents
AI doesn't read like humans
Traditional documents are designed for human eyes—layouts, fonts, page breaks. AI doesn't need any of that. What AI needs is structured context: relationships, hierarchy, and meaning. Kontext outputs Context Packets, not PDFs.
Buttons Over Chat
Don't type the same prompts
Chat interfaces make you type the same prompts over and over. 'Expand this.' 'Summarize that.' 'Give me counterarguments.' In Kontext, these are buttons. One click. Done.
Regenerate, Don't Edit
AI writes fast, let it rewrite
Why spend 10 minutes editing AI output when you can regenerate in 2 seconds? Kontext encourages regeneration over manual editing. Keep clicking until it's right.
Zero Interface
No folders, no filenames
Filing systems were designed for paper. In the AI era, everything is searchable, embeddable, and instant. Kontext has no folders, no filenames. Just click and think.
Semantic Zoom
See what matters at each level
Zoom out to see the big picture—just titles and clusters. Zoom in for details. Your canvas adapts to your current cognitive load, showing summaries or full content based on your zoom level.
AI-Native Nodes
Built for machine understanding
Every node in Kontext carries metadata that AI understands: relationships, types, constraints. When you connect nodes, you're not just drawing lines—you're building a knowledge graph that AI can traverse.
Five Axioms
The beliefs that underpin everything we build.
The unit of work is shifting from 'document' to 'context packet'.
Documents were containers for human consumption. Context packets are structured data for AI agents.
The best interface for AI is not chat.
Chat is linear, ephemeral, and repetitive. Visual, spatial interfaces preserve structure and enable non-linear thinking.
Humans should direct, AI should execute.
Your job is to provide direction: expand this, challenge that, go deeper here. AI handles the heavy lifting.
Context flows through connections.
When you connect nodes, context cascades. Upstream nodes inform downstream AI operations automatically.
The future is multi-agent, not single-model.
Different AI models excel at different tasks. Kontext lets you use the right model for each operation.
The Vision
We envision a future where thinking and AI execution are seamless. You pour your thoughts onto an infinite canvas. AI presets shape and refine them. The output is a Context Packet that any AI agent can immediately act upon—no copy-paste, no re-explaining, no context loss.
Kontext is the bridge between human cognition and AI capability.