One AI toolkit forbusiness analysts
Turn elicitation notes into requirements. Legacy screens into specs. An analysis session your AI tools can read. One menu-bar app, one shortcut.
A requirement is a thread
of decisions, not a paragraph
What you clarified in the 10am workshop has to still be attached when you draft acceptance criteria at 2pm. Khint keeps an analysis session on your Mac and hands its compacted context to every agent you run, so the draft already knows what was agreed: the legacy screen you captured at nine, the requirement you structured at eleven, and the page you publish after lunch are all one thread.
Start a session before a task. Every capture, every run and every note you drop lands in it, in order, so the thread of what you were doing survives the interruptions.
Premium €9.99 → €12.99 · live May 15 · CS owns the FAQ
Jira
Confluence
Notion
Linear- Read active sessionMemory
- Search your KnowledgeKnowledge
- Reply to emailAgent
The reply lands here.
The real Memory and Knowledge surfaces. Both live on your machine, never on our servers.
Two ways to turn a workshop into a spec
Both start from the same rough notes. They differ on what the draft is allowed to know, and on how much of yesterday you have to restate.
The decision is in yesterday's doc
The tools are all fine on their own. What none of them keeps is the thread between the workshop, the legacy screen and the page you publish after lunch.
- You restate the module, the stakeholder and the constraint at the top of every prompt, because nothing you typed this morning is still in scope.
- The old admin screen you have to replace has no documentation, so the field names get retyped by hand or guessed at.
- How the team defined that term in March is somewhere in Confluence, and finding it is a tab hunt you do instead of writing.
- The spec comes back inside a chat tab, so publishing it is a separate job with its own copy-paste and its own formatting fixes.
Khint, with the thread still attached
One analysis session holds the pass. Every agent you run reads its summary, so the draft already knows what the workshop agreed.
- An analysis session carries the workshop into the acceptance criteria. You stop restating the scope at every prompt.
- Drag a box over the legacy screen and the labels and field names come back as text, or question the screenshot instead of retyping the UI.
- Index a tool in Knowledge and every run quietly pulls the matching excerpts out of your own Jira and Confluence.
- Khint drafts from what you heard, never from nothing. You keep authorship and review; what disappears is the blank page.
The four pieces
an analysis pass runs on
A BA needs the thread kept, the corpus searchable, the draft shaped and the undocumented screen readable. These are the four pages behind that, and each is a shipped feature rather than a roadmap slide.
Memory
One session per analysis pass. What was agreed at ten is still attached when you draft acceptance criteria at two.
Knowledge
Your Jira, Confluence, Notion and Linear indexed on this Mac. A draft can cite what your team already wrote down.
Agents
Raw notes to a problem statement, the functional requirements and the open questions, in your format rather than a generic one.
Capture
The legacy admin screen nobody documented. Drag a box for the labels and field names, or question the screenshot directly.
What changes across an analysis pass
Three things, and none of them is a new place to keep your notes.
- StopsRe-establishing the context at every draftThe session carries what the stakeholder agreed in the workshop into the criteria you write after lunch, so you stop reopening your own notes to remind yourself.
- LandsThe spec, on the page your reviewers openWith Atlassian connected the palette creates the Confluence page from the refined requirements. One page per run, so every publish stays explicit.
- Gets backThe blank page at the top of every specYou're editing a draft built from what you actually heard rather than starting one. Authorship and review stay yours; the empty document is what goes.
Common questions from business analysts
What can a business analyst actually do with Khint?
Select any text (workshop notes, a stakeholder email, a Slack thread), hit Cmd+Shift+K, and run an AI agent on it in place: turn rough notes into a structured requirement, draft acceptance criteria, or tighten a spec paragraph. The same palette captures any region of your screen to text, pulls and creates Jira issues, and creates Confluence pages once you connect an Atlassian API token. You work where the requirement already lives, not in a separate chat tab.
Can Khint write a requirements document or BRD for me?
Khint drafts from what you give it, not from nothing. Select your raw elicitation notes and run a reusable agent: it returns a clean problem statement, functional requirements, and open questions in your format. You keep authorship and review; Khint removes the blank-page work. When the section is ready, create a Confluence page from the palette so the doc lands where your stakeholders read it.
How does Khint help with requirements traceability?
Start a Memory session for an analysis pass. Everything you capture, pull, and run gets compacted into a short running context, so the decision you clarified at 10am is still attached when you draft acceptance criteria at 2pm. The session is a local record of the thread of decisions behind a requirement, and Claude Desktop or Codex can read the same live context over the bundled MCP server.
Does my requirements data leave my machine?
Only what you explicitly act on. An agent sees the text you selected plus, optionally, the active Memory session's compact summary, nothing else. Memory sessions live in a local SQLite file on your Mac and are never synced to a server. Atlassian and Linear credentials stay in your OS keychain.
Is there a free plan?
Yes. The free tier includes 300 credits a month, about 10 AI actions a day, no credit card. Paid plans start at €7/month: Pro at €29/month gives 4,000 credits a month, and every plan includes all the work integrations. Khint runs on macOS 13+ and Windows 10 22H2+.
Run your next elicitation with it
Free with 300 credits a month, about 10 AI actions a day. No credit card. Connect Atlassian when you're ready: it takes an API token, not an admin.