Konnectify MCP Servers

Expose Konnectify tools, workflows, and integrations to AI clients through a secure, dedicated MCP endpoint — with full control over authentication, tool availability, and activity monitoring.

Dedicated MCP endpoints OAuth 2.0 authentication Custom AI tools Activity monitoring

What is an MCP Server?

An MCP Server acts as a bridge between Konnectify and AI assistants. It exposes selected tools through a standardised MCP endpoint, allowing AI clients to discover and invoke those tools through natural language.

By exposing tools through MCP, AI assistants can move beyond answering questions and actively perform actions on behalf of users.

AI clients can

Discover available tools

Understand tool descriptions and inputs

Execute actions and workflows

Retrieve results from tool executions

Interact with business systems through natural language

Why use MCP Servers?

Enable AI-powered actions

Allow AI assistants to execute business actions through tools exposed from Konnectify.

"Create a Freshdesk ticket and notify the support team in Slack."

Reuse existing automations

Expose existing Konnectify tools and workflows without rebuilding them for AI use cases — AI assistants leverage the same business logic already in use across your organisation.

Build custom AI tools

Create purpose-built tools designed specifically for AI interactions.

Customer support Sales operations Reporting Data retrieval

Secure client access

Control which AI clients can access your MCP Server using OAuth 2.0 authentication with full activity monitoring and logs.

Creating an MCP Server

Create a dedicated MCP Server in three steps from the left navigation.

Open MCP Servers Navigate

From the left navigation menu, click MCP Servers. This opens the MCP Server management page.

Create a new MCP Server Configure

Click Create MCP Server and provide the following details.

Field

Description

Name

Name of the MCP Server

Slug

Unique identifier used in the endpoint URL — cannot be changed after creation

Description

Optional description of the server

Authentication

OAuth 2.0 (recommended) or None

OAuth 2.0

Recommended

OAuth PKCE authentication for production environments. Securely authorizes AI clients before access is granted.

None

Testing only

Allows clients to connect without authentication. Use only in testing or controlled environments.

Important

The server slug becomes part of the MCP Endpoint URL and cannot be modified after creation. Example: https://your-domain.konnectifyapp.co/mcp/server/support-tools

Create the server Final step

Click Create MCP Server. Konnectify generates a unique MCP Endpoint URL that can be used by external AI clients.

MCP Endpoint URL format

https://your-domain.konnectifyapp.co/mcp/server/server-name

Adding tools to an MCP Server

After creating the MCP Server, navigate to the Configure tab. Under Configured Tools, click Add Tool. You can expose tools in two ways.

Add existing tools Reuse

Select from available Konnectify tools and app actions already in your workspace. These become immediately accessible to connected AI clients.

App actions

Workflow actions

Integration actions

Existing business automations

Create custom tools Custom

Click Create Tool to build a purpose-built tool designed for AI interactions. Define each aspect of the tool's behaviour.

Tool name and description

Inputs and outputs

Business logic and execution flow

Configuring OAuth clients

If OAuth 2.0 authentication is enabled, navigate to the OAuth Clients tab and click Create Client.

Field

Description

Client name

A friendly name for the client — e.g. Claude

Redirect URI

Callback URL required by the MCP client — e.g. https://claude.ai/api/mcp/auth_callback. Enter one URI per line for multiple values.

Allowed scopes

Permissions the client can request — e.g. mcp:access

Token auth method

client_secret_post or client_secret_basic — select the method required by your MCP client

Important

The client secret is displayed only once during client creation. Copy and store it securely before closing the dialog.

Monitoring activity

Navigate to the Activity tab to monitor MCP Server usage and troubleshoot issues.

Tool name

Which tool was invoked

Execution status

Success or failure

Total executions

Success and failure counts

Duration & timestamps

Execution time and timing

Common use cases

MCP Servers are suited to any scenario where AI assistants need to interact with business systems through structured, controlled tool access.

AI-powered customer support

Create, update, and retrieve support tickets

Trigger support workflows

CRM operations

Create leads and update contacts

Trigger sales workflows

Internal business processes

Expose internal workflows as reusable AI tools — triggered through natural language for approvals, reporting, and operations.

Reporting and data retrieval

Allow AI assistants to fetch information from connected systems and present actionable insights.

MCP Servers vs MCP Global

Konnectify provides two ways to expose capabilities through MCP. Choose based on how much control and separation you need.

MCP Servers

Dedicated endpoints

Create separate MCP endpoints with their own tools, authentication settings, OAuth clients, and activity logs. Best suited for teams that require distinct tool collections, environments, or client-specific access.

Separate tool collections per server

Per-server OAuth clients

Isolated activity logs

MCP Global

Single organisation endpoint

A single organisation-wide MCP endpoint exposing all available Konnectify capabilities through one centralised connection. Ideal for organisations that want a simplified MCP setup without managing individual servers.

One URL, all capabilities

No per-server management required

Fastest path to connection

Best practices

Use OAuth 2.0 whenever possible

Use clear, descriptive tool names and detailed descriptions so AI clients can select the right tool reliably

Expose only the tools required for your specific use case — avoid unnecessary tool surface area

Review activity logs regularly to monitor usage and catch issues early

Store OAuth client secrets securely — the secret is shown only once during creation

Test tools in a non-production environment before exposing them to live AI clients

Get started

Connect AI assistants to your workflows.

Create an MCP Server, expose your tools, and let Claude, ChatGPT, or Cursor drive your automations through natural language.

Get started free