Skip to main content

Documentation Index

Fetch the complete documentation index at: https://getkontext.io/docs/llms.txt

Use this file to discover all available pages before exploring further.

Features and Releases are how Kontext understands what your Product is made of and how it evolves over time. Together, they allow the analysis pipeline to anchor detected Problems to specific parts of your Product — and keep that anchoring accurate as your Product changes.

​
Features

A Feature tree is a hierarchical description of everything your Product is made of. It can be as broad or as granular as you need — from top-level product areas down to specific capabilities.

​
Why Features matter

When Kontext’s AI analysis pipeline processes incoming Feedback, it uses your Feature tree to narrow down Problems to specific areas of your Product. This is what makes the difference between knowing that your users have a Problem and knowing where in your Product the Problem lives. Without Features defined, Kontext will still detect Problems and attach them to Contexts and Actors, but you won’t be able to pinpoint which part of your Product is affected. For product managers deciding what to build next, that specificity is essential.

​
Defining your Feature tree

Your Feature tree doesn’t need to describe every detail of your Product. Focus on the areas you care about tracking — the things where knowing “there’s a Problem here” would change your priorities. A few tips:
  • Think in terms of user-facing capabilities, not technical components
  • Go deeper where you expect friction — if onboarding is a known pain point, break it into sub-Features
  • Stay high-level where things are stable — not every area needs granular tracking
  • Use an LLM to draft an initial tree from your product docs or marketing site, then refine from there

​
Watching Features

You can subscribe to any Feature using the watch button. When new Problems are detected against a watched Feature, you’ll be notified — so you can stay on top of the areas you care most about without monitoring everything.

​
Releases

A Release is a marker in time that captures how your Product evolved. While Features describe what your Product is, Releases describe when it changed.

​
What a Release does

When you create a Release, you declare a point in time when your Product’s Feature tree changed. A Release can:
  • Introduce new Features
  • Update existing Features
  • Deprecate Features that are being phased out
  • Remove Features that no longer exist

​
How Releases affect the pipeline

The Feedback analysis pipeline always operates against the latest Release. When new Feedback comes in, Kontext uses the Feature tree as it stands in the most recent Release to anchor Problems.
There is no notion of Feedback per Release — all incoming Feedback attaches to the latest Release. If your Product has different versions deployed per customer, that’s not something Kontext supports at this stage.

​
When to create Releases

Create a Release whenever your Product evolves in a way that changes the Feature tree. This could be a shipping milestone, a major update, or any moment where the structure of what your Product offers changes. The key is that each Release keeps the pipeline in sync with the current reality of your Product.

​
Automatic Releases via GitHub

If you use the GitHub Integration, Kontext watches for GitHub releases and automatically creates a corresponding Kontext Release whenever one is published. This keeps your Feature tree in sync with your shipping cadence without any manual steps.