kanban-tui
MCP server for kanban-tui, a terminal-based task manager, that exposes task, board, and column management commands for AI agents to interact with.
README
kanban-tui
A customizable terminal-based task manager powered by [Textual][textual] with multiple backends. Now also usable in co-op mode with AI agents (check the CLI interface and MCP Server section for more info).
Demo

Try kanban-tui instantly without installation:
uvx kanban-tui demo
Features
<details><summary>following the xdg basedir convention</summary>
kanban-tui utilizes [pydantic-settings] and [xdg-base-dirs] user_config_dir to save
the config file and user_data_dir for the sqlite database.
you can get an overview of all file locations with uvx kanban-tui info
</details>
<details><summary>Customizable Board</summary>
kanban-tui comes with four default columns
(Ready, Doing, Done, Archive) but can be customized to your needs.
More columns can be created via the Settings-Tab. Also the visibility and order of columns can be adjusted.
Deletion of existing columns is only possible, if no task is present in the column you want to delete.
</details>
<details><summary>Multiple Backends</summary>
kanban-tui currently supports three backends.
- sqlite (default) | Supports all features of
kanban-tui - jira | Connect to your jira instance via api key and query tasks via jql. Columns are defined by task transitions.
- claude | Read the
.jsonfiles under~/.claude/tasks/. Boards are created for each session ID. Supports only a subset of features
</details>
<details><summary>Multi Board Support</summary>
With version v0.4.0 kanban-tui allows the creation of multiple boards.
Use B on the Kanban Board-Tab to get an overview over all Boards including
the amount of columns, tasks and the earliest Due Date.
</details>
<details><summary>Task Management</summary>
When on the Kanban Board-Tab you can create (n), edit (e), delete (d), move between columns (H, L), or reorder within a column (J down / K up).
Movement between columns and reordering within a column are also supported via mouse drag and drop.
Task dependencies can be defined, which restrict movement to the Doing (start_column).
To have the restrictions available, the status columns must be defined on the settings screen.
</details>
<details><summary>Task Dependencies</summary>
Tasks can have dependencies on other tasks, creating a workflow where certain tasks must be completed before others can proceed.
- Add Dependencies: When editing a task, use the dependency selector dropdown to add other tasks as dependencies
- Remove Dependencies: Select a dependency in the table and press enter to remove it
- Blocking Prevention: Tasks with unfinished dependencies cannot be moved to start/finish columns
- Circular Detection: The system prevents circular dependencies (Task A depends on Task B, Task B depends on Task A)
- Visual Indicators: Task cards show visual cues for dependency status:
- 🔒 "Blocked by X unfinished tasks" - Task has dependencies that aren't finished yet
- ❗ "Blocking Y tasks" - Other tasks depend on this one
- ✅ "No dependencies" - Task has no dependency relationships
- CLI Support: Dependencies can be managed via the CLI with the
--depends-onflag when creating tasks, or using the--forceflag to override blocking when moving tasks </details>
<details><summary>Database Information</summary> The current database schema looks as follows. The Audit table is filled automatically based on triggers.
erDiagram
tasks }|--o| categories: have
tasks }|--|| audits: updates
tasks ||--o{ task_dependencies: "blocks"
tasks ||--o{ task_dependencies: "blocked_by"
tasks {
INTEGER task_id PK
INTEGER column FK
INTEGER category FK
TEXT title
TEXT description
DATETIME creation_date
DATETIME start_date
DATETIME finish_date
DATETIME due_date
TEXT metadata
}
task_dependencies {
INTEGER dependency_id PK
INTEGER task_id FK
INTEGER depends_on_task_id FK
}
boards }|--o{ columns: contains
boards }|--|| audits: updates
boards {
INTEGER board_id PK
INTEGER reset_column FK
INTEGER start_column FK
INTEGER finish_column FK
TEXT name
TEXT icon
DATETIME creation_date
}
columns ||--|{ tasks: contains
columns }|--|| audits: updates
columns {
INTEGER column_id PK
INTEGER board_id FK
TEXT name
BOOLEAN visible
INTEGER position
}
categories {
INTEGER category_id PK
TEXT name
TEXT color
}
audits {
INTEGER event_id PK
DATETIME event_timestamp
TEXT event_type
TEXT object_type
INTEGER object_id
TEXT object_field
TEXT value_old
TEXT value_new
}
</details>
<details><summary>Visual Summary and Audit Table</summary>
To give you an overview over the amount of tasks you created, started or finished, kanban-tui
provides an Overview-Tab to show you a bar-chart on a monthly, weekly or daily scale.
It also can be changed to a stacked bar chart per category.
This feature is powered by the [plotext] library with help of [textual-plotext].
There is also an audit table, which tracks the creation/update/deletion of tasks/boards and columns.
</details>
Installation
You can install kanban-tui with one of the following options:
uv tool install kanban-tui
pipx install kanban-tui
# not recommended
pip install kanban-tui
I recommend using [pipx] or [uv] to install CLI Tools into an isolated environment.
To be able to use kanban-tui in your browser with the --web-flag, the optional dependency
textual-serve is needed. You can add this to kanban-tui by installing the optional web-dependency
with the installer of your choice, for example with [uv]:
uv tool install 'kanban-tui[web]'
Usage
kanban-tui now also supports the kanban-tui entrypoint besides ktui.
This was added to support easier installation via [uv]'s uvx command.
Normal Mode
Start kanban-tui with by just running the tool without any command. The application can be closed by pressing ctrl+q.
Pass the --web flag and follow the shown link to open kanban-tui in your browser.
ktui
Demo Mode
Creates a temporary Config and Database which is populated with example Tasks to play around.
Kanban-Tui will delete the temporary Config and Database after closing the application.
Pass the --clean flag to start with an empty demo app.
Pass the --keep flag to tell kanban-tui not to delete the temporary Database and Config.
Pass the --web flag and follow the shown link to open kanban-tui in your browser.
ktui demo
Clear Database and Configuration
If you want to start with a fresh database and configuration file, you can use this command to delete your current database and configuration file.
ktui clear
Create or Update Agent SKILL.md File
With version v0.11.0 kanban-tui offers a CLI Interface to manage tasks, boards and columns. This is targeted mainly for agentic use e.g. via [Claude][claude-code], because references will be made only by ids, but some commands are also ergonomic for human use (e.g. task or board creation).
ktui skill init/update/delete
CLI Interface to manage Tasks
The commands to manage tasks, boards and columns via the CLI are all build up similarly. For detailed overview of arguments
and options please use the --help command.
Note that not every functionality is supported yet (e.g. category management, column customisation).
ktui task list/create/update/move/delete
ktui board list/create/delete/activate
ktui column list
MCP Server
In addition to skills, kanban-tui can be run as a local mcp server, which exposes the ktui task/board/column commands.
This requires the optional mcp dependency, which can be installed via uv tool install kanban-tui[mcp]. It utilizes [pycli-mcp]
to directly expose the commands.
Using the bare ktui mcp command shows the instruction to add kanban-tui mcp to [claude-code]. The server itself is
started using the --start-server flag.
ktui mcp
Show Location of Data, Config and Skill Files
kanban-tui follows the [XDG] basedir-spec and uses the [xdg-base-dirs] package to get the locations for data and config files.
You can use this command to check where the files are located, that kanban-tui creates on your system.
ktui info
Feedback and Issues
Feel free to reach out and share your feedback, or open an [Issue], if something doesn't work as expected. Also check the [Changelog] for new updates.
<!-- Repo Links --> [Changelog]: https://github.com/Zaloog/kanban-tui/blob/main/CHANGELOG.md [Issue]: https://github.com/Zaloog/kanban-tui/issues
<!-- external Links Python --> [textual]: https://textual.textualize.io [pipx]: https://github.com/pypa/pipx [PyPi]: https://pypi.org/project/kanban-tui/ [plotext]: https://github.com/piccolomo/plotext [textual-plotext]: https://github.com/Textualize/textual-plotext [xdg-base-dirs]: https://github.com/srstevenson/xdg-base-dirs [pydantic-settings]: https://pypi.org/project/pydantic-settings/ [pycli-mcp]: https://github.com/ofek/pycli-mcp
<!-- external Links Others --> [XDG]: https://specifications.freedesktop.org/basedir-spec/basedir-spec-latest.html [uv]: https://docs.astral.sh/uv [claude-code]: https://code.claude.com/docs/en/overview
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.