Kensa - Overview

Kensa is a desktop IDE for QA test cases. Your test cases live as plain Markdown files inside your own git repository, and a built-in terminal sits right next to the editor so you can drive local AI coding agents - Claude Code, Codex, Aider - against those files without leaving the app.

---

#The core idea

Most test-management systems (TestRail, Zephyr, Xray, Qase) are hosted web apps. Your cases live in someone else's database, behind an API, versioned by nobody. Kensa flips that:

• Every test case is a .md file with YAML frontmatter. Readable in any editor, greppable, diff-friendly.
• Your whole project is a folder you put under git. Pull requests review test cases. History, blame, branches, and backups come for free.
• There is no server and no cloud. Everything runs locally on your machine. Kensa never phones home and stores no credentials.
• AI is brought by you, not bundled. Kensa gives you a real terminal with the current case wired into the environment. You run your own already-paid-for CLI agent there. No second AI subscription, no API keys handed to a vendor.

The result is a workspace where writing, editing, reviewing, and maintaining test cases feels like working in a code repository - because it is one.

---

#Who it's for

Primary audience: technical QA engineers, AQA / SDET, and QA leads who:

• already run a CLI AI agent (Claude Code, Codex, …) under their own account,
• are comfortable in a terminal,
• feel the friction of a heavy hosted TMS, and
• work in a team with a git-based workflow.

The product is built for people who want their test cases to live alongside their code, get reviewed in PRs, and be editable by the same AI agents they already use to write code.

---

#Why people switch to it

---

#How it works in one paragraph

You open (or create) a project folder. Kensa reads the suites/ tree and shows your cases in a sidebar. You click a case and edit it in a structured editor - title, priority, status, tags, preconditions, steps, attachments, description. Saves write back to the .md file. A git panel lets you stage, commit, branch, push, and review diffs without leaving the window. A terminal panel runs your shell with the active case exported as $TMS_CASE, so you can type claude "review $TMS_CASE and add negative cases" and the agent edits the file in place. When the agent prints a draft, you right-click the terminal selection and drop it straight into the editor.

---

#What's included (feature pillars)

• Test case editor - structured, single-screen editing with auto-save.
• Suite tree - hierarchical folders of cases, drag-and-drop, multi-select.
• Bulk actions - change priority/status/tags/suite for many cases at once.
• Search - fuzzy quick-search (Cmd+K) and full-text extended search.
• Schema & fields - customizable, versioned case structure with a split-screen designer.
• Source control - a GitHub-Desktop-class git panel inside the IDE.
• Terminal & AI - real PTY with the current case wired in, plus one-click setup of the Kensa AI agent plugin.
• Import / Export - TestRail, Qase, Allure, universal CSV - both directions.
• Test plans, shared steps, runs - reusable case groupings, a shared-step library, and manual execution tracking.
• Tools - HTTP runner, mobile command palette, regex tester, JWT decoder, JSON diff, timestamp converter.
• Dashboard - project health at a glance.
• kensa CLI - a 30-command binary that ships with the app and is also usable standalone and by AI agents.

Each pillar has its own page under features/.

---

#Platforms

• Windows 10 / 11 (x64)
• macOS 12+ (Intel + Apple Silicon)
• Linux Ubuntu 22.04+ / Fedora 38+ (x64, best-effort)

Built on Tauri 2 (Rust core + native WebView), so the app is small and fast and ships as a normal desktop installer per platform, with auto-update built in.

---

#What it is not

• Not a hosted/cloud service - there is no Kensa server.
• Not a bundled AI product - you bring your own agent.
• Not a production-code editor or an autotest runner - it manages test cases, not test code.
• Single-project per window today (multi-project is on the roadmap).