The Model Context Protocol (MCP) is an open standard that lets AI tools talk to each other's data and capabilities. Taskade runs a hosted MCP server, so external AI clients like Claude Desktop, Cursor, Windsurf, and VS Code can connect straight into your workspace and work with your projects, tasks, and agents.
TL;DR: MCP is "USB-C for AI." The hosted Taskade MCP server is on every paid plan: point Claude Desktop, Cursor, or VS Code at your workspace. The opposite direction runs through automations, where the MCP Client connector calls tools on any remote MCP server. Try MCP →
The Four Layers of Workspace DNA
What is MCP?
The Model Context Protocol (MCP) is an open, standardized protocol that provides a consistent way for AI models to discover and interact with external tools, APIs, and data sources. Think of it as a "USB-C for AI": one universal connector that works with everything.
Instead of building custom integrations for each service, MCP provides a single protocol that any tool can implement. This means:
- Universal connectivity: One protocol connects to any MCP-compatible service
- Standardized tool discovery: An AI client automatically discovers what tools a server offers and how to use them
- Vendor-agnostic: Works across AI providers and tool ecosystems
- Growing ecosystem: Hundreds of MCP servers available for popular services
How MCP Works in Taskade
MCP describes a two-way street, and it helps to keep the two directions straight:
1. Taskade as an MCP server
This is the direction Taskade ships. Taskade runs a hosted MCP server, and any MCP-compatible client can connect to it:
- Available on every paid plan: Starter and up, no separate add-on
- Zero setup: Nothing to deploy or maintain on your side
- Secure sign-in: Connect with your Taskade API key
- Works with any MCP client: Claude Desktop, Cursor, Windsurf, VS Code, and more
The setup steps further down this page walk through connecting your first client.
2. Taskade as an MCP client
The opposite direction runs through your automations. The MCP Client connector holds a remote MCP server's address, lists the tools that server publishes, and calls any of them as a step. It works on every plan, with your plan setting how many automation runs you get.
An AI agent cannot reach an outside MCP server on its own. Let it make the judgment call, then hand its answer to an automation step that carries the call. The same automations cover 100+ integrations for every service that does not speak MCP.
The Integrations Directory
Taskade's Integrations Directory provides 100+ ready-to-use integrations for the automation engine. Browse them at taskade.com/integrations.
Native Integrations (31)
These are deeply integrated with Taskade's automation engine and provide the richest experience:
| Category | Integrations |
|---|---|
| Communication | Slack (12 capabilities), Discord (6), Telegram Bot (11), Microsoft Teams (3), WhatsApp Business (3), Twilio (1) |
| Gmail (4), MailChimp (1) | |
| Google Workspace | Google Sheets (8), Google Drive (18), Google Calendar (4), Google Docs (4), Google Forms (1) |
| CRM & Sales | HubSpot (5), Apollo (1) |
| E-Commerce | Shopify (13), Stripe (20) |
| Social Media | LinkedIn (2), Twitter/X (2), Facebook Pages (2), Reddit (2), YouTube (2) |
| Developer | GitHub (11), HTTP Request (1) |
| Forms & CMS | Typeform (1), Webflow (1), WordPress (1) |
| Scheduling | Calendly (3), Schedule (1) |
| Content | RSS (2), Zoom (3) |
The Wider MCP Ecosystem
The protocol is vendor-neutral and the ecosystem is growing fast, with servers published for database systems, cloud platforms, development tools, knowledge bases, payment platforms, and analytics tools. Because Taskade speaks the same standard on the server side, any client that can reach those servers can reach your Taskade workspace the same way.
Connect Taskade to Tools
Automations are how Taskade reads from and writes to other services today. The flow below shows a form submission enriching a record and landing in Slack, all through native integrations:
Step 1: Open Connections
Click Connections in the left sidebar of your workspace. The area has two tabs, Connections and API.
Step 2: Pick an Integration
- Browse the Integrations Directory
- Click the service you want to add
- Follow the OAuth or API key setup flow
- The integration is now available in the automation builder
Step 3: Build the Flow
Add a trigger, then chain actions from the connected services. A typical flow looks like this:
- Trigger: New form submission
- Action 1: Look up the customer in your CRM
- Action 2: Enrich the record with an AI agent
- Action 3: Send a Slack notification
Built-in Tools with AI Agents
Your Custom AI Agents come with a large built-in toolkit, including web search, code execution, and file analysis. Agents choose the right tool on their own during a conversation, then fold the result into their answer. For anything outside that set, hand the work to an automation the agent can trigger.
Agent tools are part of Taskade's Tools for AI Agents system. All tool usage follows your agent's permissions and knowledge training.
For detailed API documentation, see the Taskade Developer API guide.
MCP FAQ
Quick answers on plans, credits, Taskade Genesis apps, and security. For more questions, see the Frequently Asked Questions accordion further down this page.
Which plans include MCP?
Every paid plan, from Starter up. There is no separate MCP add-on. See pricing for current details.
Do MCP calls cost credits?
Work an external client kicks off in your workspace uses your plan's AI credits, the same as if you had run it in Taskade. There are no separate MCP fees.
Can AI tools reach my apps?
An MCP client connects to your workspace, so it can reach the projects, tasks, and agents behind your Taskade Genesis apps. The published app itself stays a normal web app for your visitors.
What can an MCP client do?
The Taskade MCP server exposes tools for browsing your workspaces, inspecting a project's contents, creating projects, editing the tasks inside an existing project, managing and prompting your agents, and writing app files back. Your client picks from that set the same way it picks any other MCP tool.
Is MCP secure?
Yes. Connections run over encrypted HTTPS and authenticate with your Taskade API key, so a client only ever reaches the workspaces that key can see. Revoke the key at any time to cut off access.
Ready to try it? Create a workspace, then point Claude Desktop or Cursor at it with the steps below.
Connect Claude or Cursor
Taskade works as an MCP server, meaning external AI tools like Claude Desktop, Cursor, Windsurf, VS Code, and any MCP-compatible assistant can connect to your Taskade workspace and use its capabilities. This is available on every paid plan.
What external AI tools can do with your Taskade workspace:
- Manage projects, tasks, and databases
- Create and chat with AI agents
- Run automation workflows
- Access workspace data and media
Set Up: Claude Desktop
Step 1: Install the Taskade MCP server. Open your terminal and run:
npx -y @taskade/mcp-server
Step 2: Add Taskade to your Claude Desktop configuration. Open claude_desktop_config.json and add:
{ "mcpServers": { "taskade": { "command": "npx", "args": ["-y", "@taskade/mcp-server"], "env": { "TASKADE_API_KEY": "your-api-key-here" } } }}
Step 3: Get your Taskade API key from taskade.com/settings/api.
Step 4: Restart Claude Desktop. You should see Taskade listed in the available MCP tools.
Set Up: Cursor
In Cursor, go to Settings → MCP Servers and add:
{ "taskade": { "command": "npx", "args": ["-y", "@taskade/mcp-server"], "env": { "TASKADE_API_KEY": "your-api-key-here" } }}
The same configuration works for Windsurf, VS Code (with Copilot MCP support), and any other MCP-compatible tool.
Open source: The Taskade MCP server is available at github.com/taskade/mcp. You can also use it to generate MCP servers from any OpenAPI spec.
Frequently Asked Questions
What does the Taskade MCP server let my AI tool do?
It lets an MCP client like Claude Desktop or Cursor work with your Taskade workspace directly. The client can browse your workspaces, inspect what is inside a project, create projects, add and update the tasks in a project you already have, prompt your agents, and write app files back, all without leaving the tool you are already in. You stay in your editor or chat window while your Taskade data comes to you.
Which plan do I need for the Taskade MCP server?
The hosted Taskade MCP server is available on every paid plan, from Starter up. That covers using Taskade from an external AI tool such as Claude Desktop, Cursor, Windsurf, or VS Code. See pricing for current details.
Can Taskade connect out to other MCP servers?
Yes, through automations. The MCP Client connector points an automation at any remote MCP server and calls its tools as steps, on every plan. An agent cannot reach a server on its own, so let it hand the work to an automation. For services that do not speak MCP, automations cover 100+ integrations including Slack, Google Drive, GitHub, HubSpot, and Stripe.
What is the difference between built-in agent tools and MCP?
Built-in tools are the large native toolkit that ships with every AI agent, like web search, code, and file analysis. They run inside Taskade. MCP is the open protocol that lets a separate AI client, such as Claude Desktop, reach into your Taskade workspace from the outside. The two solve different problems: built-in tools give your agents capabilities, while MCP gives your other AI tools access to your Taskade data.
How do I connect an AI tool to Taskade over MCP?
Run npx -y @taskade/mcp-server, add the small config block to your client (Claude Desktop, Cursor, Windsurf, or VS Code), then paste in an API key from taskade.com/settings/api. Restart the client and Taskade appears in its list of MCP tools. The full steps for each client are above.
Can an AI tool reach my Taskade Genesis apps over MCP?
An MCP client connects to your workspace, so it can reach the projects, tasks, and agents that sit behind your Taskade Genesis apps. The published app itself stays a normal web app for your visitors. If you want a published app to talk to outside services, wire that up with automations.
Does MCP work with every AI model?
MCP is model-agnostic by design. Any client that speaks the protocol can connect to Taskade, whichever model that client happens to be running. Inside Taskade, your agents keep their choice of 15+ frontier models from OpenAI, Anthropic, and Google plus leading open-weight providers.
Related guides
- Tools for AI Agents. The built-in agent tools.
- Custom AI Agents. Build agents with their own instructions and knowledge.
- Automations. Connect Taskade to outside services today.
- Taskade API. Another way to call Taskade from external code.
- MCP Developer Docs. Client setup, advanced usage, and connector details.
- Integrations Directory. 100+ native integrations for automations.