> ## Documentation Index
> Fetch the complete documentation index at: https://docs.kanaeru.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Glossary

> Kanaeru terms in one place.

<Tip>
  In the web app, turn on **Reading helper** in your Profile to underline these terms wherever they appear.
</Tip>

| Term | Definition |
| - | - |
| **Autonomy dial** | The control that sets how far Kana may go without checking with you first, separately for each kind of work |
| **Blueprint** | A generated mockup, diagram, or spec showing what Kana understood, produced before any code is written so you can confirm it matches what you meant |
| **Context source** | A repository, issue, document, or thread that Kana reads for context, attached to a workspace, outcome, or session |
| **Delivery review** | The end-of-outcome check where you go through what was actually built, and the evidence for it, before accepting it |
| **Evidence package** | The audit trail behind an outcome: test results, code diffs, deployment logs, approvals, and gate results |
| **Kana** | The Kanaeru teammate you work with in chat, and the name of the web app |
| **Kana session** | One working session of a given type (vision, refinement, planning, review, task planning, task execution, retrospective). Each turn can run on a different agent machine, runtime, or teammate, and the session keeps one shared history |
| **Koa** | Kana's operator agent. He runs an outcome's whole lifecycle in Kana for you, one Koa per outcome, and escalates to your human inbox only when a call is yours |
| **Agent machine** | A computer running the Agent Manager, where agent sessions execute |
| **Outcome** | One cycle of work toward your vision that ends in something you can actually use. Kanaeru's unit of delivered value |
| **Pipeline** | The ordered set of steps that delivers one task, from implementation through checks to delivery |
| **Quality gate** | A checkpoint work must pass before it can move on, such as tests, review, or build checks |
| **Refinement** | A short working session that improves or reorganizes a vision's backlog |
| **Requirement** | A WHEN … SHALL … statement a task must satisfy |
| **Roadmap** | Every item needed to reach a vision, grouped into outcomes |
| **Runtime** | The coding agent a session runs, such as Claude Code or Codex |
| **Spec drift** | A place where what was built no longer matches the approved spec or blueprint, or where a task's approved content was edited and the docs need to catch up. Listed in the **Specification Drifts** panel; drifts don't block phase changes |
| **Task** | A unit of work within an outcome, with its own pipeline, worktree, and pull request |
| **Understanding gap** | A place where a blueprint reveals that what Kana understood doesn't match what you meant. Caught before building, not after |
| **Verified outcome** | An outcome that passed every quality gate and that you explicitly accepted |
| **Vision** | The long-term result you're aiming for. One vision usually takes several outcomes to reach |
| **Workspace** | Where a vision and everything under it lives in the web app |


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.