Use kanman without a tracker | kanman documentation

You do not need Jira, GitHub Issues or GitLab issues to work with kanman. An AI team can keep its stories on its own kanman board: columns, stories with acceptance criteria and a Demonstrate step, comments and history all live in the team. You can connect a tracker later and link the stories you already have.

Choose the kanman board

When you set up an AI team, the first step lists kanman board (no external tracker) next to your Jira, GitHub and GitLab connections. It is there even when the workspace has no connection yet. Pick it, and the team’s stories live on the team’s own board.

The Columns step then shows the columns of your kanman board. Tell kanman what each one means:

Role	Meaning
Backlog	Stories that exist but are not ready to build
Refining	Stories kanman is clarifying
Ready	Stories that should be worked on next, by a developer or by kanman
In progress	Someone is working on the story
Review	The change was checked and waits for a human
Done	Finished

Each story has its own page with the acceptance criteria, the Demonstrate step, and a Comments tab with comments and history.

Repositories are optional

With the kanman board, the Repositories step is optional:

Without a repository, developers take Ready stories from their own Claude Code and kanman needs no git connection. Your code stays in your checkout. See Use kanman from Claude Code.
With repositories and a connection to your code host (GitHub or GitLab), kanman can also work on Ready stories itself, in addition to developers working from Claude Code.

You can add repositories later in the team settings under Repositories.

What changes without a repository

A team without a repository starts on the Trial preset. kanman has no repository to run an acceptance spec in, so no acceptance spec runs, and the evidence says so plainly:

No acceptance spec was run, so the outcome is not proven. Review the change carefully.

kanman still reviews the submitted change against the acceptance criteria and the team’s policy limits. To prove outcomes with an acceptance spec, add a repository with an acceptance manifest.

The dry run

There is no tracker to read, so the Dry run step does not analyze open issues. Instead it explains how stories reach your kanman board: through Intake, by moving a story to Ready, and by a developer or kanman picking it up.

How stories get onto the board

Stories come from Intake:

Open the team’s Intake page and describe a requirement in your own words.
kanman drafts stories, each with acceptance criteria and a Demonstrate step. Edit or reject drafts as needed; Write requirements kanman can finish helps with the wording.
Click Add N stories to the board, for example Add 3 stories to the board. The stories land in Backlog. Their keys come from the initials of the project name, for example PAYME-1 for a project called “Pay Me”.
Move a story to Ready when it should be worked on.

Where the evidence goes

When a change was checked, kanman posts the evidence as its own comment on the story. Open the story page and the Comments tab to read it, next to your team’s comments and the history of the story.

Model access

kanman needs model access to draft and review stories. Either:

add your own provider key under Settings > AI providers (Anthropic, OpenAI, Azure OpenAI or Amazon Bedrock) and choose Use your own key for the team, or
use kanman’s billing where it is offered for your team (Pay through kanman).

What providers receive is described in What model providers see. Plans: https://kanman.ai/en/#pricing.

Connect a tracker later

When your company adopts Jira, GitHub Issues or GitLab issues, you can move the team over without losing anything:

Open the team’s settings, Tracker. It shows kanman is your tracker.
Click Connect a tracker. Owners and admins can do this.
Pick a tracker connection and the project. If the workspace has no tracker connection yet, add one first: see Connect GitHub or Connect Jira and GitLab.
Map the tracker’s statuses to the team’s columns. kanman starts from your current columns.
Click Connect. kanman then offers Link existing stories: it proposes matching open issues by the same key, the same title or a similar title. Confirm the pairs you want to link.

Nothing is duplicated. Linked stories follow the tracker from then on: title, description and status come from the issue. Stories you leave unlinked stay kanman stories on the board.