Kanban MCP
This server is a headless Kanban MCP server backed by SQLite, exposing tools to manage projects, boards, tickets, epics, metrics, documentation links, run history, and shared staging leases.
Project management: list, create, and fetch project metadata (key, name, description, repo path, automation settings).
Board & column operations: get full board with columns/tickets/epics, create/rename/delete columns, set WIP limits and ordering.
Ticket operations: create, update, move, delete, reorder, set estimates, search/filter by column/priority, find by ID, and get top ticket in a column.
Completion workflow: moving tickets to Done requires a completion summary; supports deployment proof and structured completion evidence.
Epics: list, create, update, and delete epics with name, description, and color.
Metrics & Scrum: fetch burn-up, burndown, and velocity metrics; list allowed Scrum estimate values.
Documentation links: link docs to tickets, list docs for a ticket, and find tickets linked to a doc path.
Run history & batches: inspect ticket run history (events/artifacts) and batch details with planned ticket items.
Staging lease coordination: inspect, acquire, assert, heartbeat, and release a shared staging lease with fencing tokens (no credentials or hosted connections included).
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Kanban MCP create a ticket to fix the login button on the staging board"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Kanban MCP
Built by Natanael Mota and used to organise the development of Vortex Files.
This is the headless MCP edition of the active Kanban MCP TypeScript implementation. It provides a local MCP server backed by SQLite. The companion web UI is intentionally not included. Projects and tickets are stored only on the machine running the server.
Sample project
These are real screenshots of the companion interface running Launchpad, a fictional AI research-assistant project in a separate, empty database. No private project or customer data is shown.
Board overview: tasks grouped into Backlog, In Progress, Review and Done, with estimates and epics.

Ticket detail: an example handover task with acceptance criteria and a next action for the following agent session.

The screenshots show my companion web UI, including its separate co-pilot controls. This repository currently contains the headless MCP server, not that UI or its co-pilot integration.
Related MCP server: Local Agent Automation MCP
Built while building Vortex Files
I used Kanban MCP to organise the work of building Vortex Files, my file-sharing and client-portal product. It gave my coding agents a shared record of tasks, progress and what to work on next across sessions.
This repository shares the headless MCP server, without my private project data or companion web interface.
Requirements
Node.js 22 or newer and npm
An MCP client supporting local stdio servers
Install
npm ci
npm testStart the server with npm start, or configure your MCP client as described in SETUP.md. A first run creates a new, empty database under ~/.kanban-mcp/kanban.db. Set KANBAN_DATA_DIR to use a different writable directory. Existing databases are never bundled with this release.
Features
The MCP interface supports project listing/creation, board and ticket lookup, ticket search, project metrics, column and ticket create/update/move/delete, ordering and Scrum estimates, epics, documentation links, ticket run-history, and batch inspection. Tickets moved to Done require a completion summary. Staging-lease tools coordinate shared staging/UAT operations; use them only when you intentionally manage a shared test environment. They do not contain staging credentials or connect to any hosted environment by themselves. See tool descriptions in the MCP client for parameters and behavior.
Data and privacy
This is a local process, not a hosted service. Its stdio MCP transport has no authentication; connect only to clients and processes you trust. The included source code has no bundled customer board or credentials. Back up your database before replacing it.
License
MIT, copyright Natanael Mota. Retain the copyright and permission notice in copies or substantial portions of the software. No visible in-app credit is required. See LICENSE.
Available Tools
32 toolskanban_column_top_ticketGet Column Top TicketC
Return the highest-priority ticket (lowest priority value) from a column.
Inputs: projectKey (required), columnId, skipDone, columnName, +1 more. Returns: result
kanban_create_columnCreate ColumnC
Add a new column to a project
Inputs: name (required), projectKey (required), wipLimit, orderIndex. Returns: column
kanban_create_epicCreate EpicC
Create a new epic in a project
Inputs: name (required), projectKey (required), color, description. Returns: epic
kanban_create_projectCreate ProjectB
Create a new project with optional custom column names
Inputs: key (required), name (required), columns, description. Returns: project
kanban_create_ticketCreate TicketC
Create a new ticket in a project column
Inputs: title (required), projectKey (required), epicId, columnId, +4 more. Returns: ticket
kanban_delete_columnDelete ColumnA
Delete a column by id (removes contained tickets)
Inputs: columnId (required). Returns: success
kanban_delete_epicDelete EpicC
Delete an epic by id
Inputs: epicId (required). Returns: success
kanban_delete_ticketDelete TicketC
Delete a ticket by id
Inputs: ticketId (required). Returns: success
kanban_find_column_ticketsFind Tickets in ColumnA
Filter tickets within a column by priority band. Supply priorities or priorityMin/priorityMax bounds.
Inputs: projectKey (required), limit, columnId, columnName, +3 more. Returns: result
kanban_find_ticketFind Ticket by IDA
Look up a single ticket with project, column, and epic context. Returns full ticket details including completionSummary for completed tickets.
Inputs: ticketId (required). Returns: result
kanban_get_batchGet BatchC
Fetch a Kanban batch with its planned ticket items.
Inputs: batchId (required). Returns: batch
kanban_get_boardGet BoardA
Fetch the full board for a project, including columns and tickets. All tickets include completionSummary field (populated for completed tickets).
Inputs: projectKey (required). Returns: board, scrumValues
kanban_get_metricsGet Project MetricsB
Fetch burn-up, burndown, and velocity metrics for a project
Inputs: projectKey (required). Returns: metrics
kanban_get_ticket_runsGet Ticket RunsA
Fetch append-only run history for a ticket, including events and artifacts for each run.
Inputs: ticketId (required). Returns: runs, ticketId
kanban_link_docLink Documentation to TicketC
Associate a documentation path with a ticket.
Inputs: docPath (required), ticketId (required), linkType, projectKey, +2 more. Returns: link
kanban_list_doc_linksList Docs Linked to TicketB
Retrieve documentation links associated with a ticket.
Inputs: ticketId (required). Returns: links
kanban_list_epicsList EpicsA
List all epics for a project
Inputs: projectKey (required). Returns: epics
kanban_list_projectsList ProjectsA
List available projects
Inputs: no named parameters. Returns: projects
kanban_list_scrum_valuesList Scrum EstimatesA
Return the allowed Scrum poker estimate values
Inputs: no named parameters. Returns: values
kanban_list_ticket_linksList Tickets Linked to DocB
Find all tickets referencing a documentation path.
Inputs: docPath (required), projectKey (required). Returns: links, docPath
kanban_move_ticketMove TicketA
Move a ticket to another column. When moving to Done, provide completionSummary with: 1) What was done, 2) Future improvements, 3) Separated work (links to related tickets). Deployment proof is optional.
Inputs: ticketId (required), columnId, priority, columnName, +3 more. Returns: ticket
kanban_rename_columnRename ColumnB
Rename or update metadata for a column by id
Inputs: columnId (required), name, wipLimit, orderIndex. Returns: column
kanban_reorder_column_ticketsReorder Column TicketsA
Set the vertical order for tickets in a column. First ticket in the list receives priority 0.
Inputs: columnId (required), ticketIds (required). Returns: tickets
kanban_search_ticketsSearch TicketsC
Search tickets in a project with optional filters and pagination.
Inputs: projectKey (required), limit, query, offset, +5 more. Returns: result
kanban_set_estimateSet Ticket EstimateC
Assign a Scrum poker estimate to a ticket
Inputs: estimate (required), ticketId (required). Returns: ticket
kanban_staging_lease_acquireAcquire Staging LeaseA
Atomically acquire the one shared staging lease for a ticket candidate. Build the immutable preview before calling this. If another holder is active, acquired is false and no token is returned.
Inputs: ownerId (required), ticketId (required), projectKey (required), baselineRef (required), +3 more. Returns: lease, acquired, leaseToken, retryAfterSeconds
kanban_staging_lease_assertAssert Staging LeaseC
Validate the holder token and fencing generation immediately before a staging alias change, migration, shared-fixture mutation, acceptance, or rollback.
Inputs: leaseToken (required), projectKey (required), fencingToken (required). Returns: lease, reason, success
kanban_staging_lease_heartbeatHeartbeat Staging LeaseA
Extend an active staging lease while targeted hosted UAT is still running. An expired or superseded holder cannot renew.
Inputs: leaseToken (required), projectKey (required), fencingToken (required), ttlSeconds. Returns: lease, reason, success
kanban_staging_lease_releaseRelease Staging LeaseA
Release the staging lease after the candidate is accepted or the previous immutable deployment is restored. A stale fencing generation cannot release a successor lease.
Inputs: leaseToken (required), projectKey (required), fencingToken (required). Returns: lease, reason, success
kanban_staging_lease_statusGet Staging Lease StatusA
Inspect the shared staging lease without revealing its secret holder token.
Inputs: projectKey (required). Returns: lease
kanban_update_epicUpdate EpicC
Update epic name, description, or color
Inputs: epicId (required), name, color, description. Returns: epic
kanban_update_ticketUpdate TicketA
Update ticket fields such as title, description, column, estimate, epic, or completion summary. When moving a ticket to Done, provide completionSummary with: 1) What was done, 2) Future improvements, 3) Separated work (links to related tickets). Deployment proof is optional.
Inputs: ticketId (required), title, epicId, status, +8 more. Returns: ticket
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 32 tools
Most tools map cleanly to distinct resource-action pairs, but kanban_update_ticket and kanban_move_ticket both allow moving tickets and updating completion summaries, which creates possible confusion. The two doc-link tools also have inverse orientations that could be mistaken for each other.
Names follow a mostly consistent kanban_verb_noun pattern, but kanban_column_top_ticket lacks a verb and find/search are used interchangeably. Overall, the convention is predictable and readable.
With 32 tools, the surface is heavy and exceeds the comfortable range for an agent to quickly select between operations. Many tools are individually reasonable, but the count feels bloated for a kanban server, especially with niche entries like kanban_list_scrum_values and kanban_get_batch.
Core CRUD/lifecycle coverage exists for projects, columns, tickets, epics, doc links, and the staging lease. Notable gaps include no project update/delete, no doc-link removal, and no batch mutators, but agents can work around most of these.
Maintenance
Related MCP Connectors
Project tracker for AI agents and teams: file, claim, update and hand off tickets over MCP.
Local-first task manager: create, edit, and complete tasks, projects, and checklists via MCP.
Minimal issue tracker: projects, objectives and trackable plans, written by agents over MCP.
Project management shared by people and AI agents, with persistent project state through MCP.
Related MCP Servers
AlicenseNot gradedqualityBmaintenanceEnables AI agents to create, query, update, claim, hand off, and complete tasks on a local SQLite-backed, project-namespaced kanban board through stdio MCP tools, including filtering for ready/blocked/stale work and exchanging task attachments. It also lets agents list projects and manage statuses, ownership, and evidence without any server or accounts.MIT- AlicenseNot gradedqualityCmaintenanceEnables MCP clients to manage SQLite-backed TODOs and schedule durable coding-agent jobs over stdio, with persistent job memory, run history, and a read-only dashboard.MIT
- AlicenseNot gradedqualityBmaintenanceExposes a local-first task board over stdio JSON-RPC so MCP-capable agents can list and create projects and tasks, atomically claim the next ready task with a lease, heartbeat, log progress, manage dependency graphs, and mark tasks done or failed with results. All state lives in a single local SQLite file, so no shell command assembly or external services are needed.MIT
- AlicenseNot gradedqualityCmaintenanceExposes a single Streamable HTTP MCP endpoint that lets agent clients create, move, and search tasks on project boards with buckets and labels, and create and edit nested markdown documents that can be linked to cards. All writes are actor-tagged and audit-logged, with signed webhook events so fleets can coordinate without echo loops.Apache 2.0