Product Documentation
  • đź‘‹Welcome to K3 Labs
  • Getting Started
    • Quick Start Guide
  • Automate
    • How to create a Workflow
    • Design Page: Key Sections
    • Trigger Set Up
      • Scheduled / Manual Trigger
      • Smart Contract-based Trigger
        • Token Smart Contracts Trigger Examples
        • Swap/Staking Smart Contracts Examples
        • NFT Collection Smart Contracts Examples
      • Wallet event-based
    • Functions Overview
      • Details on Read functions datapoints
      • Transform Function Details
        • Transform Custom Formula
      • JSON Data Preview
      • Custom Inputs
    • What are Actions?
  • AI Automation Suite
    • Intro
    • AI Agent
      • Web3 Data Fetcher Tool (powered by Uniblock)
      • MCP Server
  • Choosing AI Model
  • Deploy
    • Deploy Page
    • Project Creation
  • Integrations
    • Coinbase
    • Slack
    • Telegram
    • Cubist Wallet
    • Safe Wallet
    • Hyperliquid
  • Stellar Blockchain
    • Deploying & Writing to a Stellar Smart Contract (Oracle Example)
  • Tech Documentation
    • K3 Framework
      • Abstract
      • K3 Consensus
    • Networks & Usage
    • K3 Wasm
      • K3 Wasm Rust SDK
      • K3 Wasm Internal Specifications
    • Operator onboarding
Powered by GitBook
On this page
  1. AI Automation Suite
  2. AI Agent

MCP Server

What is an AI Tool?

AI tools in K3 Labs let your AI agents move beyond simple prompts. With tool integrations, your agents can fetch live data, interact with third-party APIs, and automate complex actions—making your workflows smarter and more powerful.


What is Model Context Protocol (MCP)?

Model Context Protocol (MCP) is a new open standard—originating from Anthropic—for connecting large language models to external tools, APIs, and data sources. MCP is rapidly becoming the default for advanced AI integrations across the industry. It provides a universal way for AI agents to communicate with a growing ecosystem of services, making it possible to automate almost anything, simply by plugging in an endpoint.

  • Open standard: Widely adopted across leading AI and automation platforms.

  • Plug-and-play: Instantly connect to any tool or service supporting MCP, no custom code required.

  • Future-proof: As more providers adopt MCP, your options expand instantly.


What Can You Do With MCP?

With the MCP Server tool in K3 Labs, your AI agent can:

  • Query external APIs and databases—just enter the endpoint and, if needed, your authentication.

  • Fetch and analyze blockchain data or off-chain analytics in real time.

  • Trigger actions or workflows on external platforms, all via standard protocol.

  • Some popular use cases include: Google Sheets, Gmail, Stripe, SQL Databases (e.g., PostgreSQL), Browsing, Cloud Platform, etc.

And much more - any API or tool that supports MCP can be plugged into your workflow.

Model Reasoning Differences When connecting external tools via MCP, remember: advanced models (like Claude Sonnet 4.0, Gemini 2.5 Pro, or GPT-4o) can chain together multi-step actions, troubleshoot failed requests, and automatically retry or rephrase queries as needed. Simpler models will only perform the exact action described in your prompt, one time. Your integration experience will depend on the model’s reasoning power.


  1. Add the MCP Server Tool:

    • In your AI Agent function, select MCP Server under AI Tool.

    • Enter the endpoint URL.

    • Enter your API key for authentication.

  2. Write a Prompt: Example: Give me top 10 tokens traded on uniswap by volume

  3. Run the Workflow: Your AI agent calls The Graph via MCP, processes the data, and uses it in your automation.


Why Use MCP in K3 Labs?

  • Maximum flexibility: Connect to hundreds of popular tools (Sheets, databases, payments, cloud, email, and more).

  • No custom code: Just configure and go—MCP handles the communication.

  • Always expanding: As the ecosystem grows, so do your automation possibilities.


With MCP, K3 Labs makes it easy to bring external data, analytics, and actions into any blockchain workflow. You can automate, orchestrate, and scale—all in one place.

PreviousWeb3 Data Fetcher Tool (powered by Uniblock)NextChoosing AI Model

Last updated 3 days ago

Example: Connecting as MCP

Get the MCP Endpoint: Example:

Get Your API Key: and copy your API key.

The Graph’s Token API
https://token-api.thegraph.com/sse
Sign up with The Graph