VibeKit MCP
This server lets you manage VibeKit cloud apps end-to-end: deploy, host, edit, monitor, and automate them from any MCP client.
Deploy and host apps — create apps from starter templates or deploy GitHub repos, then list, get details, redeploy, roll back deploys, view logs, restart/stop/start, manage env vars, and delete apps.
Chat with app AI agents — send build/change requests to an app's coding agent, monitor its live progress, and read chat history.
Manage databases — enable Postgres for an app, check database status, inspect schema, run read-only SQL queries, and browse tables with pagination.
Run QA — trigger automated headless-browser QA tests and check results.
Automate coding tasks — submit autonomous coding tasks, get status/results, wait for completion, cancel tasks, and create recurring cron schedules.
Manage your account — check plan, credits, and usage, and discover/fetch implementation skills.
Enables deploying GitHub repositories to VibeKit hosting, submitting coding tasks that commit code to GitHub, and managing app deployments sourced from GitHub repos.
This server has a remotely hosted endpoint. Open its connector listing for connection details, authentication requirements, and setup instructions:
vibekit-mcp
MCP server for VibeKit, deploy apps, manage hosting, and chat with AI agents from any MCP client.
This package is for VibeKit cloud/API access. It does not connect your local Claude Code instance to Telegram. For local-machine remote control, use vibekit-agent.
Use it remotely (no install)
VibeKit also runs as a hosted remote server, so clients that accept a remote MCP URL (claude.ai web connectors, ChatGPT, etc.) need no install:
https://mcp.vibekit.bot/mcp?api_key=vk_your_api_key_herePaste that URL into your client's "custom connector" / "remote MCP server" field. Clients that let you set headers can instead send Authorization: Bearer vk_your_api_key_here and use the bare https://mcp.vibekit.bot/mcp. See Get an API key below.
Prefer a local stdio install (e.g. Claude Desktop)? Use the steps below.
Related MCP server: vibekit-mcp
Get an API key
Keys start with vk_. Two ways to get one:
Instant, no signup: one request returns a key.
curl -X POST https://vibekit.bot/api/v1/auth/register # Response: {"apiKey": "vk_...", "plan": "free", "credits": 0}Save the returned key and reuse it. Each call to
/auth/registercreates a brand-new account, so call it once, not per request.From your account: if you already use VibeKit, copy or regenerate your key in the web dashboard or the iOS app settings.
Installation
npm install -g vibekit-mcpSetup
Get a VibeKit API key (see Get an API key above).
Add to your MCP client config (e.g. Claude Desktop) (
~/Library/Application Support/Claude/claude_desktop_config.jsonon macOS):
{
"mcpServers": {
"vibekit": {
"command": "vibekit-mcp",
"env": {
"VIBEKIT_API_KEY": "vk_your_api_key_here"
}
}
}
}Restart your MCP client
Available Tools
Hosting
Tool | Description |
| List all hosted apps |
| Get details about a specific app |
| List starter templates for |
| Create new app from a template |
| Deploy GitHub repo to hosting |
| Redeploy app with latest code |
| List an app's recent deploys with status and commit |
| Roll an app back to a previous deploy |
| Get application logs |
| Restart an app |
| Stop an app |
| Start a stopped app |
| Get app environment variables |
| Set app environment variables |
| Delete an app permanently |
Agent
Tool | Description |
| Chat with an app's AI agent |
| Get agent status |
| Get chat history with agent |
Database
Tool | Description |
| Enable a Postgres database for an app |
| Get database status and connection info |
| Get the database schema (every table and its columns) |
| Run a read-only SQL query (SELECT only, up to 200 rows) |
| Browse one table's rows with pagination and sorting |
QA
Tool | Description |
| Run automated QA tests |
| Get QA test results |
Tasks
Tool | Description |
| Submit a coding task |
| Get task status/result |
| List recent tasks |
| Wait for task completion |
| Cancel a running task |
| Create recurring scheduled task |
| List scheduled tasks |
| Delete scheduled task |
Account
Tool | Description |
| Get account info (plan, credits, usage) |
| List implementation skills |
| Fetch specific skill content |
Example Usage
Once configured, you can use prompts like:
"Deploy my GitHub repo to VibeKit and create a new app"
"Chat with the AI agent for my app about adding a contact form"
"Show me the logs for my app and restart it if there are errors"
"Enable a database for my app, then show me its schema and query the users table"
"List my app's recent deploys and roll back to the last working one"
"Run QA tests on my deployed app"
"Check my VibeKit account balance and list my apps"
"Create a weekly schedule to improve my app's performance"
Environment Variables
Variable | Description | Default |
| Your VibeKit API key (required) | none |
| API base URL |
|
Related Packages
vibekit-cli: terminal client for VibeKit cloud workflowsvibekit-agent: Telegram bridge for local Claude Code on your own machine
Links
Dashboard (view or regenerate your API key)
Privacy Policy
VibeKit's privacy policy: https://vibekit.bot/privacy
What this MCP server does with your data, concretely:
Every tool call is forwarded to the VibeKit API (
https://vibekit.bot/api/v1) over HTTPS, authenticated with the API key you configure. Tool inputs and outputs are processed by VibeKit to run your apps and agents, under the privacy policy above.The server itself keeps no state: nothing is written to disk, no analytics or telemetry are collected by this package, and your API key is read from configuration (
VIBEKIT_API_KEY/ the extension's key field) and sent only tovibekit.bot.Data retention, third-party sharing, and deletion are governed by the policy above. Questions: support@vibekit.bot.
Available Tools
2 toolsvibekit_database_statusVibeKit Database StatusARead-onlyIdempotentInspect
Get an app's database status and connection info (whether it's provisioned, frozen, etc.).
Inputs: appId (required). Returns: Result envelope: ok=true with data on success, ok=false with error on failure.
vibekit_get_taskVibeKit Get TaskARead-onlyInspect
Get the status and result of a task submitted via vibekit_submit_task.
Inputs: taskId (required). Returns: Result envelope: ok=true with data on success, ok=false with error on failure.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 2 tools
The two tools have completely distinct purposes: one checks database status, the other retrieves task results. No ambiguity.
Both tools use the consistent 'vibekit_' prefix followed by verb_noun pattern (database_status, get_task), showing good naming uniformity.
Only 2 tools for a server named 'VibeKit MCP' is too few, especially since a referenced tool (vibekit_submit_task) is missing, making the set feel incomplete.
The tool set is missing a critical tool (vibekit_submit_task) that is referenced in vibekit_get_task, and there are obvious gaps in domain coverage beyond database status and task retrieval.
Maintenance
Related MCP Connectors
MCP server that lets your AI assistant build and deploy apps, hosting each one at a live URL.
MCP server for your apps' tools and custom tools, plus hosted AI agents and approval-gated workflows
Talk to your public-facing AI from any MCP client — Claude, ChatGPT, Cursor, Cline, Windsurf.
Real-time chat for AI agents. Claude Code, Cursor, Cline and Codex join channels over MCP.
Related MCP Servers
- AlicenseAqualityBmaintenanceDeploy static websites, landing pages, and prototypes instantly. Free, no account needed.415790 npm6MIT
- AlicenseBqualityDmaintenanceEnables the management of AI-powered VibeKit apps, allowing users to control deployments, monitor logs, and perform database operations directly from MCP-compatible clients. It facilitates interaction with hosted AI agents and the execution of headless coding tasks through natural language.4926 npmMIT
- AlicenseAqualityBmaintenanceAn MCP server for managing Modal — apps, containers, volumes, and secrets — and for deploying & running Modal apps directly from Claude Code and other MCP clients.1285 PyPI4MIT
- FlicenseNot gradedqualityDmaintenancePublish your Claude Code setup, browse a public feed, and pull setups in. Remote MCP server + web feed in one FastAPI app.1-