How Kanso Works: From Your Website to Approved Work

Key points

• Kanso reads, it prepares, and you approve. Nothing goes public without your yes.
• Setup is a one-time-code sign-in and a few short screens: where your brand lives, what Kanso read, and four lines about your business. Each can be skipped.
• Connecting accounts is optional and one at a time. Your own Claude is the one connection that matters for drafts.
• Drafts land on the Content page. Publishing shows the exact text first, then asks for a separate confirmation.
• Verified means Kanso saw the change on the live page, not that rankings improved.

How Kanso works in three moves: it reads, it prepares, you approve

The rest of this page follows that order, screen by screen. For the wider picture, what Kanso AI is explains the whole workspace, and the documentation covers how it fits together.

Start free with your website address

You start by typing your website address on the home page and pressing Start free, and no credit card is asked for. Start free with your website address and you land on the sign-in screen.

There you enter your email and receive a one-time code. There is no password to invent, and the guided setup that follows is built to take about two minutes.

Tell Kanso where your brand lives

Kanso asks where your brand lives, and you can answer with a website, a profile or handle, a line of your own words, or several at once. One smart box takes any of them, and six tiles do the same by hand: website, Instagram, LinkedIn, X, YouTube and Reddit.

A brand can live in several places at once, and none is treated as the main one. Nothing limits how many places you add.

Each profile is read the moment you add it. Its card says “Reading”, then asks “Is this you?”. If a profile's bio links to your website, Kanso offers it with one tap and never adds it silently.

Confirm what Kanso read, then check the four lines that define your business

You confirm each profile and then check four short lines, and Kanso uses nothing until you do. The “Is this you?” card shows the bio and the recent posts as Kanso read them.

If a page refuses to be read, which happens with some Reddit and LinkedIn pages, a paste box lets you supply the text yourself. A live screen shows the profile, the latest posts as read, and a brand kit of colours, logo and style filling in, and you can carry on while it finishes.

1. What you do. One line on your business.
2. Who it is for. Your customer in a sentence.
3. What you sell. Your products or services.
4. How you sound. Your voice.

One button saves the four lines. Kanso then shows a summary of what it learned, with two optional next steps, connecting Claude and setting a goal. Positioning, things never to claim, competitors and goals can be refined later in your business documents.

What does the workspace show on day one?

On day one the workspace opens on Home, which shows a short briefing from the Chief of Growth, your team of specialists and what is waiting for you. The briefing is one sentence about what happened, what needs your yes and what comes next.

Below it sits the roster of eleven specialists, each with a status and a count of what is waiting, and the Chief itself, which you can ask questions. It answers from what Kanso knows about your business. The other screens are one click away.

• Today: a ranked list of what to do.
• Team: each specialist's own pages.
• Content: every draft and article Claude or a specialist saved.
• Results: search traffic, visitors, page speed and technical checks.
• Business: your business documents.
• Guide: ready-made sentences to paste into Claude.
• Settings: sources, connected apps and appearance.

The Agent Hub describes each specialist, and the guides are short walkthroughs for the jobs you will do most.

Which apps should you connect first, and which can wait?

Connect your own Claude first if you want drafts, and leave every other app until the job in front of you needs it. Every connection is optional and has its own click.

Connecting Claude is one copied address and three steps, and a light turns green after Claude makes its first real call. You need a Claude account that supports connectors. The setup guide covers each connection in order.

Each other account uses that platform's own sign-in, so Kanso does not see your passwords, and credentials are stored encrypted. Plans are on the pricing page.

How the Today list decides what to do first

Today puts the most useful things to do first and shows the reason on every card, so you can see why an item sits where it does. It is a ranked list, not a timetable, and nothing on it starts by itself.

A card names the action, such as write an article, improve a page, fix a website issue, draft a reply, turn an article into posts or reconnect an account. It also says why it matters, roughly how many minutes it takes and when it is due.

• Start it yourself in the specialist's pages.
• Press “Do this with Claude” so your own Claude writes it.
• Dismiss it, and give a reason.

Here is an illustrative first session, built only from the steps on this page and not a customer story.

1. A founder types a website address, signs in with a code, confirms two profiles and saves the four lines.
2. Home shows the briefing, and Today lists a card to write an article, with the reason and a minute estimate.
3. The founder presses “Do this with Claude”. Their own Claude reads the business documents and saves a draft.
4. The draft appears on the Content page. The founder edits one paragraph, which saves a new version, and approves it.
5. The founder copies the Markdown to hand to a developer. Nothing has gone public without a yes.

How drafts reach the Content page

Every draft, article and reply that your Claude or a specialist saves lands on the Content page. The table shows what you press and what arrives for six of the specialists.

On the Content page you can filter by status, such as draft, approved and published, and by channel. Open a draft to read it, edit it as a new version, and compare versions word by word. Only the workspace owner can approve.

When your Claude saves something, the workspace refreshes by itself and shows the last thing Claude did, with a Refresh button as a fallback. Nothing here is public yet. The Writer Agent, LinkedIn Agent, X Agent and Reddit Agent pages explain each channel's drafts.

How publishing works: a preview, then a separate confirm

Publishing from the Content page is two deliberate steps: a preview of the exact text and the account, then a separate confirmation. Only the workspace owner can approve or publish.

1. Preview. You see the exact text and the account it will go to.
2. Confirm. A separate click publishes it through your connected X, LinkedIn or Reddit account.
3. Or post it yourself. You can copy the text, or open the platform with the text ready, and post it on your own.

Nothing posts twice. A post moves from draft to posting to posted, and if the result is unclear, for example after a network error, Kanso does not retry and flags it for you to check. Your Claude can publish only if you have turned that on, and the preview step still applies.

Posting a Reddit reply from the Reddit Agent works from the reply you are looking at: Post to Reddit publishes it exactly as it appears at that moment, so read it first. The approval workflow page documents every approval step and who can take it.

A code change works differently: when the owner confirms, the Coding Agent opens a pull request on a new branch, marked as a proposal for human review, and stops. Your developer reviews and merges it on GitHub. You approve each piece and stay in control of what goes out.

How Kanso checks whether a fix is live

After a supported fix is made, Kanso looks at the live page again and checks one specific condition, such as the page now having a meta description. The result is verified, unchanged or inconclusive, and verified means Kanso saw the change on the live page.

It does not mean your rankings improved, because Kanso makes no promise about rankings or traffic. It can check only what is visible on the page, such as a title, a meta description or a canonical link.

The path runs from the SEO Agent's recommendation to a tracked task, then to a pull request, and after your developer merges it Kanso reads the pull request's state and marks the fix done. AI-readiness tasks are not marked verified.

How Kanso gets sharper over time

Kanso gets sharper in three ways, and in each one a person decides what changes. None of them promises better results, only drafts that sound more like your business.

• You edit. Any field in your business documents can be edited, each edit is saved as a revision with a visible history, and anything Kanso inferred has a Confirm button.
• Claude proposes. Your Claude can suggest corrections to what Kanso knows about your business, and nothing changes until you accept.
• Channel strategies refine. Each specialist keeps a short strategy for its channel that is refreshed as it learns. For X you can switch on learning from your own recent posts, and the LinkedIn Agent can learn from posts published through Kanso.

The prompt library has copy-and-edit prompts for the next job.

How Kanso works: frequently asked questions