Skip to main content
Back to the Learning Center
  • Intermediate
  • 12 minutes
  • No SEO knowledge needed

SEO with AI agents: let Claude Code work with real SEO data

An AI agent like Claude Code can check pages, suggest titles and add internal links. Without data, though, it only guesses. Here you connect it to Crawl Foundry, give it fixed rules with a skill, and decide what it may spend.

Let's start

In a nutshell

An agent is like a new colleague on day one.

It works fast and thoroughly but knows neither your website nor your rules. MCP gives it access to the data in Crawl Foundry, such as audit findings, rankings and Search Console numbers. The skill is its onboarding folder: it says which tool is for what, when the agent has to ask, and how it reports.

Why it matters: Without data, a language model fills gaps with numbers that only sound plausible. With data and rules, you get work you can check.

First things first

What to look out for first

  1. 1.Numbers only from tools

    Do thisThe agent names the tool and the date for every number, for example "latest stored audit from 28 September".

    Not thisThe agent estimates search volume or positions from memory.

  2. 2.Read first, then buy

    Do thisThe agent reads stored data for free. Before it buys new data, it states the cost and waits for your yes.

    Not thisThe agent starts a new crawl or a new research run for every question.

  3. 3.A person approves

    Do thisYou read titles, descriptions, links and structured data before they go live. Nothing is published without your explicit approval.

    Not thisThe agent writes fifty new pages without real value and puts them online straight away.

Step by step

How to set it up

  1. 1Step 1 of 4

    Connect Claude Code to Crawl Foundry

    Install Claude Code and run the terminal command that adds Crawl Foundry as an MCP server. Then start Claude Code, type /mcp and sign in to Crawl Foundry in your browser. There you check the workspace and the requested permissions and allow access. Every command is ready to copy further down.

    Done when: /mcp lists crawl_foundry as connected.

    Simplified sketch with sample data

  2. 2Step 2 of 4

    Install the skill

    Download SKILL.md and put it in the folder ~/.claude/skills/crawl-foundry-seo/. At startup Claude Code reads only the skill's name and description. It loads the rest when your task matches. You can also call it directly with /crawl-foundry-seo.

    Done when: When you type a slash in Claude Code, /crawl-foundry-seo shows up in the list.

    Simplified sketch with sample data

  3. 3Step 3 of 4

    Start safely

    Begin with a task that buys nothing and changes nothing. The agent selects the workspace, tells you its name and domain, and reads the stored overview. Once that checks out, move on to real tasks, for example: "Check the /pricing page and suggest two better titles."

    Done when: The agent names your workspace and backs every number with a tool and a date.

    Simplified sketch with sample data

  4. 4Step 4 of 4

    Set the budget

    The agent budget is on by default. Without approving each price, a connected agent may then spend up to €5.00 per action, €10.00 per workspace and €50.00 per organization each month on research and content creation. To control every purchase yourself at first, set your workspace to "No spending" in the "Agent budget" tab under "Agent, MCP & API". Only organization administrators can change budgets.

    Done when: The "Agent budget" tab shows the setting you want for your workspace.

    Simplified sketch with sample data

Ready to copy

Every command for the setup

Three commands, in this order. Crawl Foundry also shows the connection command under "Agent, MCP & API".

  1. 1. Install Claude Code

    Terminal · macOS, Linux, WSL
    curl -fsSL https://claude.ai/install.sh | bash
    Terminal · Windows PowerShell
    irm https://claude.ai/install.ps1 | iex

    Then start Claude Code in your project folder with the command claude and sign in on first launch. Claude Code needs a Claude subscription or an Anthropic Console account.

  2. 2. Connect Crawl Foundry

    Terminal · any system
    claude mcp add --transport http --scope user crawl_foundry "https://app.crawlfoundry.com/api/mcp"

    The command adds Crawl Foundry for all your projects. Then type /mcp in Claude Code and sign in. The commands for Codex, Cursor and VS Code are on the SEO MCP server page.

  3. 3. Install the skill

    Terminal · macOS, Linux, WSL
    mkdir -p ~/.claude/skills/crawl-foundry-seo && curl -fsSL https://crawlfoundry.com/downloads/crawl-foundry-seo/SKILL.md -o ~/.claude/skills/crawl-foundry-seo/SKILL.md
    Terminal · Windows PowerShell
    New-Item -ItemType Directory -Force "$HOME\.claude\skills\crawl-foundry-seo" | Out-Null; irm https://crawlfoundry.com/downloads/crawl-foundry-seo/SKILL.md -OutFile "$HOME\.claude\skills\crawl-foundry-seo\SKILL.md"

    The command creates the folder and downloads SKILL.md into it. You can also save the file with the download button and put it there yourself.

Common mistakes

Mistakes that happen a lot at the start

  • Telling the agent nothing more than "do my SEO".

    Better: Give it a concrete task with a URL and a goal, such as checking one page or suggesting titles for a category. The skill gives it the procedure.

  • Overlooking the default budget.

    Better: Check the agent budget under "Agent, MCP & API" before the agent starts. "No spending" is enough to begin with.

  • Accepting suggestions without checking them.

    Better: An agent with good data can still be wrong, especially with metadata. Read every suggestion before it goes live, and measure the result later.

Mini glossary

Terms you need here

AI agent
A program built on a language model that does more than answer: it uses tools to read files, run commands and query data. Claude Code is such an agent.
MCP
Model Context Protocol, an open standard that lets AI applications connect to external tools and data. Anthropic introduced it in November 2024.
Skill
A folder with a SKILL.md file: name, description and instructions for one kind of task. Since December 2025 the format has been an open standard that other tools read too.
Agent budget
The limit in Crawl Foundry up to which a connected agent may buy data without approving each price. It applies per organization, per workspace and per action.

Done

Connect your first agent to Crawl Foundry.

The agent reads your workspace's stored data for free. It buys new data only within the budget you set.

Go to the SEO MCP server

Cookie policyPrivacy policy