—SEO MCP server

The SEO MCP server that gets you backlinks, not just data.

Add it to Claude, Cursor, Codex or VS Code with one line, nothing to install. Your AI app picks directories that fit your product, fills each form in your own browser from a tested playbook, and checks every link went live.

LaunchDistro has 50 directories ready to submit to, each with a playbook tested on a real submission, and it counts a link as live only after opening the public page and reading its rel attribute.

Run this once in a terminal (or ask Claude Code to run it). It adds LaunchDistro for every project:

$claude mcp add --scope user --transport http launchdistro https://launchdistro.com/api/mcp --header "Authorization: Bearer <your-token>"
Nothing to installRuns in your own browserStops for sign-ins
—Definition

What is an SEO MCP server?

An MCP server that gives an AI app like Claude or Cursor SEO tools. Most only read data: keywords, rankings, backlinks. LaunchDistro does the work: it submits your startup to directories from tested playbooks, stops for you at sign-ins, and counts a link only once it's live.

—Works in

One server, every AI app you use.

Each guide covers where the config lives, how to check it's connected and what to ask first.

—Data, or done

Most SEO MCP servers tell you what to do. This one does it.

Search for an SEO MCP server and you'll find servers for keyword data, rankings and backlink research. They answer questions. None of them fills in a directory form, and that's the slow part of getting your first backlinks.

An SEO data MCP

  • Keyword volumes, SERPs, competitors' backlinks
  • Your agent reads the numbers and suggests what to do
  • You still do the submitting

LaunchDistro

  • Picks directories that fit your product from a tested catalogue
  • Fills each form in your browser, stops for you at the gates
  • Checks each link on the live page and again every week

They work well together: ask a data MCP which directories your competitors are listed on, then ask LaunchDistro to submit you to the ones in its catalogue.

—Workflows

What to ask your AI app.

Copy a prompt into Claude, Cursor or Codex. Each line under it is the tool your agent calls.

Product profile

Write the profile from your repo

“Read my README and landing page, write my product profile, and show it to me before you save it.”

Drafts name, tagline, descriptions and categories from what it reads, then saves them

save_product_profile(profile)
Planning

Pick directories that fit

“Plan a campaign of 10 directories that fit my product, DR 50 or higher, dofollow only.”

Filters the tested catalogue by authority, link type and category, and lists them for you

plan_campaign(target_url, count=10, min_dr=50)
Submitting

Submit, and stop for you at the gates

“Submit to the first 3 directories in the campaign. Tell me when a sign-in or CAPTCHA needs me.”

Fills each form in your browser from its tested playbook, and pauses at every human gate

queue_submission(slug) → handle_checkpoint(action)
Checking

Count only the links that went live

“Check which of my listings are live, and whether each link is dofollow or nofollow.”

Opens each public listing, finds your link and reads its rel attribute

verify_link(live_url, target_url) → monitor_link_health(campaign_id)
—The tools

13 tools, in five groups.

Your AI app reads these names and descriptions itself and picks the right one, so you never have to call them by hand.

Plan

  • match_directoriesWhich directories take your product: ready, one step away, or not a fit and why.
  • plan_submissionsAn ordered run, each directory with the reason it is where it is, saved to your dashboard.

Product profile

  • list_productsThe products on your account.
  • get_product_profileRead one: name, tagline, descriptions, links, and what forms ask that it's missing.
  • save_product_profileSave the profile your agent wrote from your repo.

Submit

  • submission_playbookA directory's tested playbook: what the form asks, the free plan's conditions, the traps, where to stop.
  • record_submissionRecord what happened: draft, submitted, needs you, rejected or skipped, with the confirmation page.

Check and report

  • verify_linkOpen the public listing, find your link and read its rel attribute.
  • verify_badgeCheck your site shows a directory's badge before you press its verify button.
  • report_directory_factTell us what a directory's page says now: a new price, a badge rule, a changed form.
  • suggest_directorySend us a directory you recommend, to be checked and given a playbook.
  • list_submissionsEvery directory for a product, with its status and what to do next.

Account

  • license_statusPlan, connected apps and submissions used.
—How it works

What happens when your AI app calls it.

Nothing to install

The server runs at launchdistro.com. Your AI app connects with one line or one URL and a token made for that app, which you can revoke on its own.

Playbooks, not guesses

Each directory has a playbook tested on a real submission: what the form asks, what the free plan really gives, where to stop. Your agent reads it before it touches the form.

Your browser, your accounts

Your agent fills the form in your own Chrome, signed in as you. At a sign-in, a CAPTCHA, a fee or a terms box it stops and asks you.

Only live links count

Once a directory approves you, the link is checked on the public page and its rel read. On Pro it is checked again every week.

What reaches our servers: your account, one token per AI app, your product profile and what your agent records. The forms are filled in your browser, so cookies, sessions and passwords never do.

SEO MCP questions, answered

An MCP server gives an AI app like Claude, Cursor or Codex tools it can call. An SEO MCP server gives it SEO tools. Most of them read data: keywords, rankings, backlinks. LaunchDistro is one that does work: it submits your startup to directories and checks the backlinks went live.

Your next 10 directory submissions are free.

Add it to Claude or Cursor, describe your product once, and watch the first form fill itself.