MCP Servers Overview
MCP (Model Context Protocol) servers allow your AI agents to securely interact with external applications and services. Instead of just answering questions, agents connected to MCP servers can read data, trigger actions, and automate workflows across the tools your team already uses.
What Are MCP Servers?
An MCP server acts as a bridge between your AI agent and an external service. Each server exposes a set of tools — specific capabilities like reading messages, creating issues, or managing files — that the agent can use during a conversation. The agent decides which tools to call based on the user's request, and the MCP server handles communication with the external service securely.
For example, connecting a Slack MCP server gives your agent the ability to search messages, post replies, and look up users — all within a conversation.
How MCP Servers Work
- An organization admin enables and configures MCP servers for the organization, including authentication credentials.
- An agent creator connects one or more MCP servers to a specific agent and selects which tools the agent can use.
- End users interact with the agent normally. When the agent needs to interact with an external service, it calls the appropriate MCP server tool behind the scenes.
Authentication is handled automatically — users authorize access once (for OAuth-based servers), and the agent uses that authorization for subsequent requests.
Supported Integrations
Authentication Types
MCP servers support several authentication methods, depending on the external service:
| Authentication Type | Best For |
|---|---|
| OAuth | Third-party services like Slack, Google, GitHub, Microsoft, Atlassian, Figma, Dropbox, Zapier |
| API Key | Services that use static API keys for authentication |
| Pass-Through Headers | Internal services that accept the same credentials you use with Devs.ai |
| None | Public endpoints that do not require authentication |
Most pre-built integrations use OAuth, where each user authorizes access to their own account. The platform handles token management, refresh, and secure storage automatically.
Getting Started
- Organization admins: Start with Configuring MCP Servers for Your Organization to enable and set up integrations.
- Agent creators: See Configuring extended capabilities to learn how to connect MCP servers to your agents.
- Per-server setup: Visit the individual server pages listed above for detailed configuration and OAuth setup instructions.
📝 Note: When MCP tools return large amounts of data, the platform automatically compresses the response to fit within the AI model's context window. For details, see MCP Payload Compression.