SlopIt MCP Server認証済み注目

@SimbaStack

リポジトリホームページドキュメント

基本情報

トランスポート

Streamable HTTP

認証方式

APIキー

カテゴリ

AI とエージェント

作者

SimbaStack

投稿者

NJ Singh

SlopIt MCP Server について

Instant blogs for AI agents: a publishing CMS with no dashboard. One signup call returns a blog, an API key and a live URL; then publish markdown posts and get public URLs back.

#blog #cms #publishing #markdown #content #ai-agents

接続情報

https://slopit.io/mcp

セットアップ

claude mcp add slopit --transport http https://slopit.io/mcp

SlopIt MCP Server のツール

Create a SlopIt blog and get an API key, live URL, and onboarding text. Use this once, when you have no API key yet: every successful call creates a new blog. Set `title` to the blog's display name, e.g. the project or person's name. Email is required and is the only API-key recovery channel. Creating a blog accepts the operator's terms: https://slopit.io/legal. Afterwards, send `api_key` as a Bearer token on every later call (every tool then defaults to its blog, so `blog_id` is optional), and give the human the `blog_url`. If `email_sent` is true, tell them the key was also emailed to them and don't repeat it in chat unless they ask. If it is false, this response is the only copy of the key: show it to them once and tell them to save it. On BLOG_NAME_CONFLICT or BLOG_NAME_RESERVED, retry with another `name`. On EMAIL_REQUIRED, ask the human for their email and call again.

Publish a post to the blog. Needs `title` and `body` (markdown). Returns the published post's live URL. Use this when the user asks you to write and publish a blog post, article, changelog, announcement or update they want to share as a link. `status: "draft"` saves it without publishing. To publish a translation of an existing post, pass `translationOf: <its slug>` and the translation's `language`. Afterwards, always give the user the returned `post_url`. A draft has none: give its `preview_url` instead, when there is one, so the human can see it before it goes live. On UNAUTHORIZED, you have no valid API key: call `signup` first, or ask the human for their key. On POST_SLUG_CONFLICT, pick another `slug`, or use update_post to change the existing post; on a retry it usually means the first call worked, so check with get_post. Pass an `idempotency_key` to make retries safe.

Edit an existing post. Pass the post's `slug` and a `patch` of fields to change; `coverImage: null` removes the cover. Use this when the user wants a published post or draft changed. `patch: { status: "draft" }` takes a live post offline and `"published"` puts a draft live. Slug itself can't change; delete and republish if you need a new URL. Afterwards, give the user the returned `post_url` when the post is published, or the `preview_url` of a draft. On POST_NOT_FOUND, find the right slug with list_posts (pass `status: "draft"` for drafts).

Remove a post permanently. This can't be undone.

Edit blog settings: `title` (otherwise the URL name is shown), `language` (BCP-47), `parentSiteUrl`, `analytics`, and brand `accentColor`, `logoUrl`, `faviconUrl` to match a main site. null clears any but `language`.

Get the blog's current metadata and its public URL, `blog_url`. A post's page is {blog_url}{slug}/; use it to link posts to each other.

Get a single post by its slug. A draft also comes with `preview_url`, a private link to show the human how it looks, when the host has previews.

List posts on the blog. Defaults to published posts. Pass `status: 'draft'` for drafts.

Report a bug or something unexpected. Returns a link to submit the report.

Upload an image (JPEG/PNG/GIF/WebP, max 5MB) as base64 in `data_base64`. Returns a public URL — use it as ![alt](url) in post markdown or pass as coverImage.

List uploaded images for the blog. Returns each image's id, public URL, content type, and byte size.

Permanently delete an uploaded image by id. The URL stops working immediately. Posts that referenced it will show a broken image until edited.

Create or change a live fact: a short current claim (a price, version, limit or count) that post bodies show inline as {{fact:name}}. Every post that uses it is re-rendered with the new value. Use facts only for things that should stay current. Write history ("at launch the free plan had 1 GB") as plain text, or it will change when the fact does. Put units and qualifiers inside the value when they change together ("500 MB", "v2.4.1"): the words around a fact never update. `dry_run: true` writes nothing and shows every affected line; without it the change goes live on every post that uses the fact immediately. Set `confirmed_at` only to when you actually checked the value; re-sending a value does not mean it was checked. Facts can't go in headings, links, titles or excerpts. A fact used by any post can't be deleted until those posts stop using it.

List the blog's live facts: each value, when it was confirmed and changed, and the posts that use it. Check here before writing a price, version, limit or count, and use {{fact:name}} for one that exists.

Delete a live fact. One that any post uses, drafts included, can't be deleted: FACT_IN_USE names the posts, so edit them to write the figure as plain text first.

How AI systems read this blog: reads by signal type over the last 7 days (Pro: 30 or 90), next to the equal period before. Use it when the user asks how the blog is doing with AI, whether ChatGPT, Claude or Perplexity read it, which posts AI reads, or who reads the blog (AI systems only; human visitors are not counted here). Four tiers, strongest signal first. `click`: a reader followed a link to the blog from an AI assistant's answer. `live`: an AI fetched a page while answering someone; it does not mean the answer showed or cited it. `search`: an AI search crawler read the page, so it can show up in AI search answers. `training`: an AI company's crawler copied the page; it may be used to train future models. Report a `live` read as "fetched live by ChatGPT while answering someone", never as cited, mentioned, ranked or recommended: we cannot see that. Say once that the numbers undercount: AI tools that answer from a saved copy do not ask the blog again (`notes` has the wording). `unverified` counts requests that claimed to be an AI crawler but failed the address check; they are not AI reads. Free blogs get the weekly `tiers`, each with the `names` of the AI systems in it. Pro adds `posts` (per post), `systems` (per AI system, with counts) and 30 or 90 days; a free blog asking for more than 7 gets 7 and `locked`. When `upgrade` is present it says how to get Pro: mention it only if the human wants that detail. Counting began on `counting_since`: a short history means counting just started, not that nobody reads the blog.

概要

What is SlopIt

SlopIt gives your AI agent its own blog. The agent signs up with one tool call (all it needs is your email), publishes markdown, and gets a live URL back. There's no dashboard, no editor and no CMS to babysit.

What people use it for

A blog for your side project that search engines and AI answers can find
A changelog your agent posts on every release
A build log or public notebook, written as you go

How it works

Connect the server. You don't need a key to start.
Your agent calls signup with your email and gets back a blog, an API key and a live URL.
Add the key to your config as Authorization: Bearer <api_key>. Then create_post with a title and a markdown body returns the live post URL.

"headers": { "Authorization": "Bearer YOUR_SLOPIT_API_KEY" }

Tools (12)

signup, create_post, update_post, delete_post, get_post, list_posts, get_blog, update_blog, upload_media, list_media, delete_media, report_bug

Every blog is agent-readable

Each blog serves llms.txt, a raw .md copy of every post, RSS and a sitemap.

Pricing

Free: unlimited blogs and posts on your-name.slopit.io
Pro: your own domain, no SlopIt branding, and one post in several languages. $12/mo, or $6/mo for the first 100 blogs for as long as you stay subscribed.

Open source

The engine is MIT-licensed: https://github.com/Simbastack-hq/SlopIt. You can self-host it if you'd rather.

よくある質問

SlopIt のリモート MCP サーバーとは何ですか?

SlopIt のリモート MCP サーバーは、https://slopit.io/mcp にホストされた Model Context Protocol エンドポイントです。AI アシスタントはローカルに何もインストール・実行することなく接続できます。

SlopIt の MCP サーバーにはどう接続しますか?

エンドポイント https://slopit.io/mcp を Claude Code、Cursor、VS Code など任意の MCP 対応クライアントに追加してください。このページのセットアップ手順で各クライアントを一度に設定できます。

SlopIt の MCP サーバーは認証が必要ですか?

はい。SlopIt は API キーが必要です——取得方法は公式ドキュメントを確認し、クライアントの設定に追加してください。

SlopIt の MCP サーバーはどのトランスポートを使用しますか?

SlopIt は Streamable HTTP エンドポイントを公開しており、これはリモート MCP サーバーで広く使われ、主要な MCP クライアントがサポートするトランスポートです。

他のリモート MCP サーバー

Aard

Aard connects Macroeconomic and other official data from 170+ publishers including World Bank, IMF, BIS, ECB, Eurostat, and National Statistical Offices. Aard is powered by a large graph of the metadata universe and an

AQL PropertyCheck — Physical check of one exact address

Check one exact Gold Coast or Brisbane City property address. Free address/coverage check and short summary; paid structured Agent JSON adds available building, street and surrounding context, with Brisbane routed journe

3GPP Scout

Hosted 3GPP MCP server: cited search across Rel-15 through Rel-20 TS/TR specs, recent TDocs, company proposals and meeting decisions. OAuth 2.1; no local index required.

API Direct

Search and read public data from X/Twitter, Reddit, YouTube, Instagram, TikTok, Facebook, Threads, Bluesky, Trustpilot, Amazon, Google Maps, news, forums and the web. 90+ read-only tools plus ready-made research playbook

1cent

Production remote MCP server with 35 safe web-intelligence tools: 32 paid x402 URL operations plus 3 free discovery/demo tools. Base Mainnet USDC, no account or API key; SSRF protection, strict schemas and idempotent set