Docket SEO

io.github.mattkerr09Published by GitHub account mattkerr09v1.3.152Updated Nov 6, 2026

SEO audits that run on your computer: score, ranked fixes and AI-search checks. Free to try.

The publisher's own server: docket

VerifiedSTDIODesktop onlyWeb Search & ScrapingData & Analytics

Overview

AI-generated overview

Docket SEO runs on-site SEO audits locally and lets an assistant crawl a site, read past reports, explain findings and compare audits.

What you can do
  • Crawl a site and get a score with every finding, affected pages and fixes
  • Retrieve the most recent kept audit, overall or for one site, without re-crawling
  • Explain a single finding in full: what is wrong, which pages, why it matters and the fix
  • Compare two kept audits of the same site for score and area changes and new or fixed problems
  • List every check Docket runs, grouped by area with ids and titles
When to use it
Useful when you want an assistant to audit a website's SEO, dig into specific issues and fixes, or track how a site's score and problems change between audits, all on your own machine.
Requirements
A local install of the Docket desktop app (macOS, Windows or Linux) and an MCP client pointed at the bundled docket command with the mcp argument over stdio. No API key or network port is needed. A licence key is required for the full findings; without one, audits return a limited free version.
Before you install
The free tier returns only the score, per-area problem counts and the single most important problem; full findings and fixes need a one-time $59 licence. Audits crawl the target site, so expect outbound requests to it. Kept audits are stored locally and can be read back by the assistant.

Installation

In SourceWeft

  1. Open Docket SEO in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

Docket MCP server

Docket is an SEO audit app that runs on your own computer. Its engine has an MCP server built in, so Claude, the ChatGPT desktop app or any MCP client can run an audit, read past audits and explain a problem with its fix.

The server is the docket command with the argument mcp. It talks over stdio, opens no network port, and needs no API key. You install Docket, then point your client at the engine.

Install Docket

Add it to your client

The engine is at:

  • Mac: /Applications/Docket.app/Contents/Resources/docket/docket
  • Windows: C:\Users\YOU\AppData\Local\Docket\docket\docket.exe
  • Linux: the docket file in the folder you unpacked

Claude Code:

claude mcp add docket -- /Applications/Docket.app/Contents/Resources/docket/docket mcp

Claude Desktop, in claude_desktop_config.json:

json
{  "mcpServers": {    "docket": {      "command": "/Applications/Docket.app/Contents/Resources/docket/docket",      "args": ["mcp"]    }  }}

ChatGPT desktop: Settings, then MCP servers, add a server, choose STDIO, and give it the same command with the argument mcp.

Any other MCP client works the same way: the command and the argument are all it needs.

Tools

ToolWhat it does
audit_siteCrawls a site and returns the score and every finding, with the pages affected and the fix. Takes url, and optionally max_pages, max_seconds and offline.
get_last_reportReturns the most recent kept audit, or the most recent of one site, without crawling again.
explain_findingExplains one problem from a kept audit in full: what is wrong, which pages, why it matters and the fix.
compare_auditsCompares two kept audits of the same site: the score change, each area's change, and which problems are new or fixed.
list_checksLists every check Docket runs, by area, with its id and title.

Free and paid

Without a licence key, audit_site returns the free version: the score, the count of problems in each area and the most important problem in full. A licence ($59 once) returns every finding and fix. Details at https://docketseo.app/how-to/audit-your-site-from-an-ai-assistant/

Source: README.md at commit 6223048

Tools

0
SourceWeft started this package once in an isolated sandbox, without credentials, to list its tools. That failed: the package type cannot run in the sandbox.

Version history

2
  1. v1.3.152LatestNov 6, 2026
  2. v1.3.150Nov 6, 2026

Score

40% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts

    Declare a license in the repository or the registry entry.

Other Docket integrations