mcp-vibekanban
MCP server that bridges AI coding assistants to Vibe Kanban, enabling task, session, and queue management without leaving the coding environment.
README
MCP server that bridges your AI coding assistant to Vibe Kanban. manage tasks, start sessions, send follow-ups, monitor execution — all without leaving Claude Code, Cursor, Amp, or whatever you're using.
VIBE_PROJECT_ID=<uuid> npx mcp-better-vibe-kanban
locks to a single project at startup so every tool call is zero-boilerplate. no project IDs in every request.
what it does
12 MCP tools and 5 MCP resources exposed over STDIO, HTTP, or Cloudflare Workers.
tools
task management:
| tool | what it does |
|---|---|
get_context |
project name, active task, workspace info |
list_tasks |
filter by status, configurable limit |
create_task |
create with @tag expansion in descriptions |
get_task |
full task details by ID |
update_task |
title, description, status — any combination |
delete_task |
irreversible delete |
session management:
| tool | what it does |
|---|---|
start_workspace_session |
spin up a session for any executor (Claude Code, Amp, Gemini, Codex, Cursor, etc.) |
list_sessions |
all sessions in a workspace |
get_session |
session details by ID |
send_message |
follow-up with auto-queue fallback if session is busy. supports MCP task primitives for async tracking |
monitoring:
| tool | what it does |
|---|---|
get_queue_status |
check if a message is queued |
cancel_queue |
drop the queued message |
resources
| URI | description |
|---|---|
vibe://tasks |
all tasks in the project |
vibe://context |
workspace context (when VIBE_WORKSPACE_ID is set) |
vibe://tasks/{taskId} |
single task details |
vibe://sessions/{sessionId} |
session details |
vibe://sessions/{sessionId}/queue |
queue status |
resources support subscriptions — polls on a configurable interval, emits updates on change via SHA-256 diffing.
install
# run directly
VIBE_PROJECT_ID=<uuid> npx mcp-better-vibe-kanban
# legacy aliases still work
npx vibe-kanban-better-mcp
npx vkb-mcp
or build from source:
git clone https://github.com/yigitkonur/mcp-better-vibe-kanban.git
cd mcp-better-vibe-kanban
pnpm install && pnpm build
requires Node.js 18+ and a running Vibe Kanban instance (default: npx vibe-kanban on port 9119).
configuration
| variable | default | description |
|---|---|---|
VIBE_PROJECT_ID |
— | required. UUID of the project to lock onto |
VIBE_REPO_ID |
auto-detected | repo UUID. auto-fetched if project has exactly one repo |
VIBE_API_URL |
http://localhost:9119 |
Vibe Kanban API base URL |
VIBE_WORKSPACE_ID |
— | enables vibe://context resource and get_context tool |
VIBE_RESOURCE_POLL_INTERVAL |
10000 |
resource subscription poll interval (ms) |
MCP_TRANSPORT |
stdio |
stdio or http |
MCP_PORT |
3000 |
port for HTTP transport |
client setup
Claude Desktop
{
"mcpServers": {
"mcp-better-vibe-kanban": {
"command": "npx",
"args": ["-y", "mcp-better-vibe-kanban"],
"env": {
"VIBE_PROJECT_ID": "your-project-uuid",
"VIBE_REPO_ID": "your-repo-uuid",
"VIBE_API_URL": "http://localhost:9119"
}
}
}
}
Cursor
same structure, drop it in your Cursor MCP config.
HTTP transport
MCP_TRANSPORT=http MCP_PORT=3000 VIBE_PROJECT_ID=<uuid> npx mcp-better-vibe-kanban
health check at GET /health, MCP endpoint at POST /mcp. each HTTP session gets its own isolated server instance.
Cloudflare Workers
wrangler deploy
uses Durable Objects with SQLite migrations. progress notifications and task primitives don't work in the Workers runtime.
transports
| transport | use case | command |
|---|---|---|
| STDIO (default) | Claude Desktop, Cursor, local clients | VIBE_PROJECT_ID=<uuid> npx mcp-better-vibe-kanban |
| HTTP | remote/multi-session | MCP_TRANSPORT=http MCP_PORT=3000 ... |
| Cloudflare Workers | edge deployment | wrangler deploy |
internals
single API client (src/api/client.ts) that uses curl via spawnSync instead of native fetch — works around an ARM64 macOS socket issue where Node.js fetch gets EHOSTUNREACH on LAN IPs. blocks the event loop, but reliable.
src/
index.ts — MCP server factory, transport setup
config.ts — env var loading, UUID validation
tools/
index.ts — all 12 tool definitions (Zod schemas + handlers)
api/
client.ts — VibeClient singleton, curl-based HTTP
types.ts — TypeScript interfaces for API responses
resources.ts — MCP resources, subscription manager with SHA-256 change detection
tasks.ts — MCP task primitive integration (experimental)
utils/
formatter.ts — response formatting
progress.ts — progress notification emitter
worker.ts — Cloudflare Workers entry point (compiled by Wrangler, excluded from tsconfig)
dev
pnpm dev # hot reload via tsx
pnpm build # compile to dist/
pnpm start # run dist/index.js
pnpm clean # remove dist/
license
MIT
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.