Raw signal in.Roadmap out.
Kano is an agentic workspace for product management. Feedback, discovery, strategy and roadmaps live as linked Markdown files in your own GitHub repo, and a swarm of agents keeps them collected, triaged, connected and honest. You make the calls.
- CaptureNotes, tickets, calls, CSVs
- TriageLabelled, deduped, themed
- ConnectLinked to opportunities
- DecideChecked against goals
- PlanNow, next, later
Animation: scattered customer signals flow from left to right through four agents and settle into three roadmap lanes named Now, Next and Later.
The whole job, from learning to planning
Twenty-four artifact types, from a scribbled note to a release, each one a typed document you edit in a block editor and every one linked to the evidence behind it.
Call with Northwind: they won't renew without SSO. Also asked twice about exporting the audit log. Marco thinks pricing page confuses mid-market.
Mid-market buys on security, not features
Activation in the first session
Priya, IT admin at a 300-person firm
Orbit ships SCIM next quarter
Now
Next
Later
Dump raw notes, pipe in tickets and calls. Agents label each item, find duplicates and suggest what it should become.
A swarm that does the legwork
Each agent is a Markdown file: instructions, a model, the tools it may use, when it wakes up and where it may write. Change any of it, or describe a new one in a sentence.
Feedback Collector
Pulls new tickets, calls and messages from Zendesk, Salesforce, Jira, Sheets and chat into the inbox.
every 2 hoursTriager
Reads your raw notes and suggests the feedback, decisions, opportunities and links hiding in them.
on every noteFeedback Rationalizer
Labels, dedupes and clusters feedback into themes, ties themes to opportunities, writes the weekly digest.
Mondays 07:00Weaver
Finds the goals, personas and initiatives a new item belongs to and links them. Doubtful links wait for you.
on new artifactInterview Synthesizer
Extracts insights with verbatim quotes, jobs and pains, and attaches them to personas and opportunities.
on new interviewMetrics Analyst
Queries PostHog and Salesforce, records each metric against its target and flags what moved.
Mondays 08:00PRD Writer
Drafts a PRD from the initiative's problem, outcome and linked evidence, as a proposal for you to review.
when askedRoadmap Steward
Spots overdue work, release slips and dependency conflicts, then drafts the stakeholder update.
Mondays 08:00Strategy Auditor
Reports goals with no work behind them, initiatives with no evidence and bets gone stale.
1st of the monthYours
“Every Friday, scan competitor changelogs and tell me what affects our bets.” Kano drafts the agent, you check its tools and scope, and it starts with a dry run.
you decide
Agents propose. Nothing strategic merges without you.
Scoped
Each agent writes only to the folders you list. Strategy, discovery and roadmap changes always arrive as proposals.
Isolated
Every run works in its own git worktree and is validated before a single file reaches your main branch.
Budgeted
Caps on tokens, tool calls, minutes and monthly spend. Each run records what it cost.
Reversible
Every run has an ID, a diff and a log. Revert one with a single command.
Talk to it from wherever you already work
Press ⌘J on any page. The master agent reads the workspace with the same tools the swarm uses, cites every artifact, and anything it changes lands in your review queue.
Add your Kano address as a connector and sign in. Your own Claude or ChatGPT plan does the thinking, so chatting with your product costs you no extra API tokens.
Every operation is a command. Script it, pipe it, or point Claude Code and Codex at the workspace folder and let them drive.
One endpoint per operation with a generated OpenAPI spec, API tokens scoped by role, and signed webhooks for n8n, Make and Zapier.
Log feedback from the thread where you heard it, and get review requests, digests and mentions back in the space. Google Chat today, Slack next.
More than a hundred operations, each generated once and exposed as a button, a REST endpoint, a CLI command and an MCP tool.
Any model, a different one per agent
Bring your own keys. Put a careful model on strategy and a cheap, fast one on labelling, or run the whole swarm on a local model. Keys sit in an encrypted vault, never in the repo.
Your product, as files in your GitHub repo
There is no database to export from. Each workspace is a folder of Markdown with typed frontmatter, synced to a repository you own. Every edit, by a person or an agent, is a commit.
--- id: INI-003 type: initiative title: SSO and SCIM provisioning status: approvedhorizon: now links: goals: [G-002] opportunities: [OPP-012] products: [PROD-001] --- ## Problem Mid-market IT teams can't approve us without SSO. See [[OPP-012]] and the 14 requests behind it.
Readable anywhere
Open the folder in an editor, in Claude Code, or on GitHub. Edits made outside Kano show up in the app within two seconds.
History for free
Who changed the goal, when, and what the roadmap looked like at the last review are all in git, with three-way merge when two people edit at once.
As many workspaces as you need
One per product, team or client, each with its own repo, agents and members. Switch between them from the sidebar.
Run it your way
Kano is self-hosted. Pick where the content lives and where the agents run, and change your mind later: the repo is the same in every setup.
- Content lives on
- Agents run on
- You work in
One always-on container runs the app, the API, MCP and the agent schedule. Your team opens it in a browser; every edit is pushed to your GitHub repo.
The repo opens soon.
The GitHub link isn't public yet. Leave your email and we'll send it the day Kano launches. The first 100 people on the list get it free.