3. What KaKeKiKoKu Is Not

Before diving into how KaKeKiKoKu works, it helps to clarify what it is not — because the concept is new and it's easy to confuse it with tools you already know.

Not a context file

Instruction files such as CLAUDE.md and AGENTS.md hold working rules and can be organized by scope. Their maintenance and loading depend on the client and your workflow. KaKeKiKoKu adds context attached to project resources, explicit relations, and synchronization checks; the comparison explains how they can coexist.

KaKeKiKoKu is a living system: it adds context to each file and folder separately, detects file changes and structural inconsistencies, and helps you and your AI review the affected knowledge. Detection alone does not prove that its content is correct. It's not a file — it's a semantic layer over your project.

Not documentation

Documentation explains how to use something. KaKeKiKoKu contexts capture what the code cannot express: why this decision was made, what alternatives were discarded, what constraints exist, what traps to avoid.

It doesn't replace your documentation — it complements it. Your documentation is for the people who read it. KaKeKiKoKu is for the people and AIs who build the project together.

Not just memory for your AI

Built-in client memory helps an AI remember you and your conversations, according to that client's design. KaKeKiKoKu keeps something else: the project's own knowledge. It lives beside your files, you shape and review it with your AI, and it stays the same whichever AI or client you use next. Persistent memory is one of its capabilities, not its purpose.

Not an AI skill

Skills teach the AI to do something new. KaKeKiKoKu doesn't teach it anything — it gives it context. An AI with KaKeKiKoKu doesn't know more, but it knows about your project. The difference between an AI that guesses and one that knows is the context it has available.

Not a project manager

KaKeKiKoKu doesn't manage tasks, sprints, or deadlines. Resources that require future attention can be marked as Pending: this is a standard KaKeKiKoKu flag stored with the resource, not a separate task system. KaKeKiKoKu's mission is to preserve knowledge — decisions, intent, constraints. That knowledge feeds whatever management tool you use, but KaKeKiKoKu doesn't replace it.

Actions on real project files

The knowledge workflow manages context without autonomously editing your work. Explicit Web actions can create, duplicate, move, rename, or delete project resources, and the quick editor saves real Markdown and text files. These actions are distinct from editing their contexts. Review their result and use backups when you need to recover an earlier state.

Not a cloud tool

KaKeKiKoKu has no servers, accounts, or telemetry. It runs entirely on your machine. Note that when your AI reads project knowledge, that data is processed by your AI provider according to their privacy policy.

So what is it?

KaKeKiKoKu Studio is a knowledge system. A project is its knowledge: you and your AI shape it, and Studio keeps it in structured resources that you can read, connect, and maintain across sessions and AI clients — true to your files. The Knowledge Engine is its technical core.

For a complete functional description, see chapter What Is KaKeKiKoKu. For installation instructions, visit the Quick Start Guide.