2 Add the connector
In your tools settings, add StealthGPT MCP as a connector and paste the URL.
Undetectable AI,
Now in Your Stack
StealthGPT integrates directly with MCP and n8n, so undetectable AI writing runs inside your existing automation workflows. No web form, no manual copy-paste.
MCP Serverhttps://www.stealthgpt.ai/api/mcp/mcp
- Long-form content that performs
- Human & Undetectable output
- Built for builders, agents & automations
Claude.ai as an example
Connect once. Takes two minutes.
Generate humanized, undetectable text without leaving Claude, Cursor, or Manus.
1 Get your API key
Grab a key from your API dashboard.
3 Type /humanize
Use it right inside the chat. No new tab, no copy-paste, ever again.
Use it inside your favorite LLM
Fully humanized AI writing, exactly where you need it.
Claude.aiLive
Claude CodeLive
CursorLive
ManusLive
n8nLive
WordPressLive
CodexComing Soon
ZapierComing SoonHow to Connect
Integration made simple so you can stay undetectable.
Claude.ai & Claude Desktop
- 1.
Go to Customize section
In the sidebar, click Customize, then Connectors. (On the web you can also open claude.ai/customize/connectors directly.)
- 2.
Click the + button, then Add custom connector
In the Connectors panel, open the + menu and choose Add custom connector.
Free to start
No plan required. Activate API access, add a card, and you're in. After that it's just $0.20 for every 1,000 words you generate. A StealthGPT web subscription is separate and not required.
Free to start$0.20 / 1,000 wordsNo subscription
Features
Call StealthGPT From Anywhere
MCP is the open standard for connecting AI apps to external tools, the USB-C of agents. Once StealthGPT is inside your MCP tool, an agent can call it mid-reasoning, hand it a draft, and get back undetectable, humanized text without any copy-pasting. That means one agent does drafting, detection, and rewriting in a single chain instead of you stitching systems together. And because it’s a standard, any MCP-compatible client can use it with zero custom integration work. Future-proof by design.
Drop It Into n8n
StealthGPT ships an official, n8n-verified node, n8n-nodes-stealthgpt, listed in the n8n integrations directory. Kick off full content-pipeline runs and humanize text directly inside your workflows. It runs on the async API under the hood, which fits n8n’s long-running execution model cleanly. Setup is just adding your StealthGPT API credential and connecting, and your token lives only in your n8n credential store. It’s never embedded in the node package.
Stealth Agent: One Request, Full Pipeline
Stealth Agent is your all purpose solution; it researches, drafts, and fact-checks when you ask it to, and only then humanizes, so programmatic output stays stable and high-quality even from loose instructions. One preset drives the whole sequence. Academic mode delivers structure, citations, and formal tone. SEO mode produces organized headers, search-aligned sections, and meta attributes, ideal for auto-generating blog posts and landing pages. Social mode targets LinkedIn or Medium with strong hooks and channel-specific writing. No more chaining four services to run one content pipeline.
Built for Volume
Both the Writer/Humanizer and Stealth Agent support asynchronous runs, so you don’t hold a persistent HTTP connection open. Fire a POST to start a run, get a runId back immediately, then poll for status or let StealthGPT hit your webhookUrl when it finishes. It’s the non-blocking pattern that keeps n8n flows and MCP servers reliable under high traffic, exactly what you want when you’re generating long-form content at scale.
Why it's different
Built as an API first. Not a wrapper around the web app.
Fully humanized
Bypasses every AI detector while sounding like your wrote it. Genuinely human in structure, tone, and variation.
Reliable async primitive
Idempotency keys prevent double charges on retries. Webhooks push results. Run IDs stay stable.
Full pipeline or step only
Use the end-to-end pipeline, or drop the humanization step into the workflow you already run.
Start with one request
Create a run, get a run_id back, collect the result when it’s ready. The example below is the whole loop. Idempotency keys, webhooks, and presets are covered in the docs. We keep the full reference there, not here.
Stealth Agent API
POST /api/stealth-agent/runs
{
"preset": "seo",
"prompt": "Write about async APIs",
"webhook_url": "https://..."
}
→ { "run_id": "run_abc123", "status": "pending" }
// poll GET /runs/{run_id} or receive a webhook
→ { "status": "completed", "result": "# SEO Article..." }Frequently Asked Questions
Still have more questions?
Learn more in our support page.
MCP (Model Context Protocol) is the open standard for connecting AI apps to external tools, the USB-C of agents. Exposing StealthGPT as an MCP tool lets any compatible client call it mid-reasoning, hand it a draft, and get back lower-detection, more human-sounding text, with zero custom integration work.
The API gives you direct access to humanization and writing endpoints. The Stealth Agent is a multi-phase pipeline that researches, drafts, fact-checks, and humanizes from a single request, driven by presets like Academic, SEO, and Social.
Claude.ai, Claude Desktop, Claude Code, Cursor, and Manus are live today, with n8n and WordPress supported through dedicated integrations. Codex and Zapier are coming soon.
Yes. StealthGPT ships an official, n8n-verified node, n8n-nodes-stealthgpt, listed in the n8n integrations directory. Add your StealthGPT API credential, connect, and run full content pipelines inside your workflows.
Both the Writer/Humanizer and the Stealth Agent support asynchronous runs. Fire a POST to start a run, get a runId back immediately, then poll for status or let StealthGPT call your webhookUrl when it finishes, so you never hold an HTTP connection open.
Presets drive the whole sequence from one request. Academic delivers structure, citations, and a formal tone. SEO produces organized headers, search-aligned sections, and meta attributes. Social targets LinkedIn or Medium with strong hooks and channel-specific voice.
Yes. Your token lives only in your n8n credential store and is never embedded in the node package, so credentials stay isolated from your workflow definitions.
Absolutely. Because StealthGPT is an MCP tool, a single agent can draft, run detection, and rewrite in one reasoning chain, instead of you stitching multiple systems together.
Grab an API key from your dashboard, add StealthGPT MCP as a connector in your tool of choice, paste the URL, and start generating humanized text right inside the chat.
Integrate StealthGPT Into Your Workflow
StealthGPT is no longer just a site you visit, it’s a tool you can call. The web app still works the way it always has, but the API, Stealth Agent, and the new MCP and n8n integrations mean the product now lives inside your agent’s reasoning chain, in your automation pipeline, and behind your webhook.