Skip to main content

MCP Tools & Servers

Nexla's MCP Tools & Servers lets you turn your Nexsets into AI-ready MCP tools, organize them into ToolSets, and deploy them as MCP Servers that any AI agent can consume. This guide introduces the core concepts behind MCP Tools & Servers and explains how each piece fits together.

What is MCP Tools & Servers?​

MCP Tools & Servers is Nexla's MCP-as-a-Service platform. It bridges your enterprise data -- represented as Nexsets in Nexla -- with AI agents and assistants by exposing that data through the Model Context Protocol (MCP). Rather than building custom integrations between each AI tool and each data source, MCP Tools & Servers provides a standardized layer that any MCP-compatible client can connect to immediately.

The platform is available in two ways: through the AI tab in the Nexla Platform UI, and programmatically via the REST API at https://api-genai.nexla.io/. Both interfaces give you full control over creating tools, assembling ToolSets, and deploying MCP Servers.

With MCP Tools & Servers, your data team can govern exactly which data is exposed, how it is accessed, and who can use it -- while your AI and engineering teams get instant, standardized access to that data through any MCP-compatible agent such as Claude Desktop, Cursor, or ChatGPT. Every tool execution is tracked with an immutable audit trail, giving you full visibility and compliance.

Key Concepts​

Tools​

A Tool is an AI-callable interface generated from a Nexset. When you create a tool, Nexla automatically generates the tool definition -- including its name, description, input schema, and execution logic -- based on the underlying Nexset's structure and metadata.

There are three kinds of tools:

  • nexset_read -- Read data from a Nexset. These tools allow AI agents to query and retrieve data.
  • nexset_action -- Perform actions against a Nexset, such as writing or updating records.
  • custom_tool -- A user-defined tool with custom logic and configuration.

Tools are versioned using semantic versioning (e.g., 1.0.0, 1.1.0, 2.0.0), so you can evolve tool behavior while maintaining backward compatibility for agents already using earlier versions.

ToolSets​

A ToolSet is a named collection of tools. ToolSets let you group related tools together -- for example, all tools related to a particular domain, project, or use case. Each ToolSet maintains a manifest that tracks which tools (and which versions of those tools) are included.

ToolSets serve as the unit of deployment: when you are ready to expose tools to AI agents, you deploy a ToolSet as an MCP Server.

MCP Servers​

An MCP Server is a deployed instance of a ToolSet. Once deployed, AI agents connect to the MCP Server using the Model Context Protocol and can discover and invoke all of the tools contained in the ToolSet.

MCP Servers support authentication via Nexla service keys and OAuth, ensuring that only authorized agents and users can access your tools.

MCP Gateway​

The MCP Gateway lets you register external MCP servers from third-party providers and bring their tools into your Nexla ToolSets. This allows you to combine Nexla-generated tools with tools from other services in a single, unified ToolSet -- giving your AI agents access to a broader set of capabilities without leaving the Nexla ecosystem.

Receipts​

Receipts provide an immutable audit trail of every tool execution. Each time an AI agent invokes a tool, Nexla records a receipt that captures what tool was called, who called it, when it ran, what inputs were provided, and what the outcome was. Receipts are essential for compliance, debugging, and monitoring how AI agents interact with your data.

How It Works​

The end-to-end workflow for MCP Tools & Servers follows five steps:

  1. Start with Nexsets -- You have data represented as Nexsets in the Nexla Platform. These Nexsets can come from any data source Nexla supports, including databases, APIs, file systems, and cloud services.
  2. Generate Tools from Nexsets -- Select one or more Nexsets and use the Tool Generator to create AI-callable tools. Nexla automatically infers the tool schema from the Nexset structure.
  3. Organize Tools into a ToolSet -- Group your tools into a ToolSet. You choose which tools to include and can mix Nexla-generated tools with external tools brought in through the MCP Gateway.
  4. Deploy the ToolSet as an MCP Server -- Export the ToolSet as an MCP Server. Nexla provisions a dedicated MCP endpoint with a unique URL that agents can connect to.
  5. AI Agents Connect and Use Your Tools -- MCP-compatible clients such as Claude Desktop, Cursor, and ChatGPT connect to the MCP Server URL, discover available tools, and invoke them as part of their workflows.

Accessing MCP Tools & Servers​

There are two ways to access MCP Tools & Servers:

Nexla Platform UI​

Open the Nexla Platform and click the AI tab in the main navigation. The AI tab provides a visual interface for creating tools, managing ToolSets, deploying MCP Servers, registering external servers via the Gateway, and reviewing audit receipts.

REST API​

The MCP Tools & Servers API is available at https://api-genai.nexla.io/. You can use the API to programmatically manage tools, ToolSets, servers, and receipts.

Authentication

Both the UI and API require authentication using a Nexla service key. See Service Keys for instructions on creating and managing service keys.

Get Started​

Explore the following guides to begin working with MCP Tools & Servers:

  • Getting Started -- Walk through the complete workflow from Nexset to connected AI agent.
  • Tools -- Learn how to create and manage tools from your Nexsets.
  • ToolSets -- Learn how to organize tools into deployable collections.
  • MCP Servers -- Learn how to deploy ToolSets and manage MCP Server instances.
  • MCP Gateway -- Register external MCP servers and bring third-party tools into your ToolSets.
  • Audit & Receipts -- Review the immutable audit trail of all tool executions.

For API documentation, see the MCP Tools & Servers API.