# gotoHuman Extension

> Add gotoHuman MCP Server as a goose Extension

- 网址：https://funcoding.ai/agents/goose/mcp/gotohuman-mcp/
- 来源：goose 官方文档原文（英文），Apache-2.0 许可，同步于 2026-10-11
- 官方原文：https://goose-docs.ai/docs/mcp/gotohuman-mcp

---
<!--   -->

This tutorial covers how to add the [gotoHuman MCP Server](https://github.com/gotohuman/gotohuman-mcp-server) as a goose extension to bring **human-in-the-loop approvals** into your AI workflows. With gotoHuman, goose can pause and request a review before continuing, perfect for blog drafts, code reviews,compliance checks, etc.  

<div class="callout callout-tip">

**Quick Install**

**goose Desktop**

[Launch the installer](goose://extension?cmd=npx&arg=-y&arg=%40gotohuman%2Fmcp-server&id=gotoHuman&name=gotoHuman&description=gotoHuman%20MCP%20server%20for%20human-in-the-loop%20approvals&env=GOTOHUMAN_API_KEY%3DgotoHuman%20API%20Key)

**goose CLI**

**Command**
```sh
 npx -y @gotohuman/mcp-server
```
**Environment Variable**
```sh
GOTOHUMAN_API_KEY: <YOUR_API_KEY>
```

</div>

## Configuration

**goose Desktop**

**goose CLI**

## Example Usage

<div class="callout callout-tip">

**Before You Start**

Log in to [app.gotohuman.com](https://app.gotohuman.com) and go to **Review Templates → Create Template**.  
Follow the steps to create a template and ensure you set a **Webhook Endpoint** (you can use [webhook.site](https://webhook.site) for testing).  

</div>

In this example, goose sends a LinkedIn post draft to gotoHuman for approval using the `n8n news to post` template that was created.

**goose Desktop**

1. Open a new session in goose Desktop

**goose CLI**

1. Open a terminal and start a new goose session:

```sh
goose session
```

### goose Prompt
```
Send this blog draft about goose to gotoHuman for review using my `n8n news to post` form. 

Include today’s date as the timestamp, these links: 
[goose Docs: https://goose-docs.ai/, gotoHuman: https://gotohuman.com/], 
summarize it as ‘Introducing goose’s integration with gotoHuman for human approvals,’ 
and here’s the draft:

goose is an open-source AI agent that runs locally on your machine, right in your terminal or IDE. 
It connects to MCP servers like gotoHuman to extend its abilities. While goose can automate tasks, 
sometimes you need a human in the loop. With gotoHuman MCP, you can add approval steps into any workflow;
from reviewing a LinkedIn post draft to signing off on a code change. 
Once approved, goose can continue where it left off. 
This makes agent workflows safer, more collaborative, and better suited for real-world teams.
```
### goose Output

<div class="callout callout-note">

**Desktop**

`─── list-forms ──────────────────────────`

Now let me search for some specific setup and usage information:

`─── Get-Form-Schema`

Let me search for more technical details and examples:

`─── Request-human-review-with-form ────────────`

Perfect! I've successfully sent your goose blog draft to gotoHuman for review using your "n8n news to post" form. Here's what was submitted:
Review Request Details

- Form Used: n8n news to post
- Review ID: 6Tp7Z0lqH7lu52MfuHZq
- Review Link: https://app.gotohuman.com/accounts/3kiXJvmf4UDBPKPi9ZNY/reviews/6Tp7Z0lqH7lu52MfuHZq

Submitted Content

- Timestamp: 2025-09-23
- Summary: Introducing goose's integration with gotoHuman for human approvals
- Links:
    - goose Docs
    - gotoHuman
- Draft: Your complete blog post about goose's integration with gotoHuman

The review request has been created and is now ready for human approval. You can access it through the review link provided above. Once the review is completed, you'll be able to proceed with publishing or making any requested changes to the blog post.

</div>

## Review in gotoHuman

Now if you open the **Review Link** from goose’s output, you’ll see the request waiting in your gotoHuman dashboard:  

![](https://funcoding.ai/official-assets/goose/documentation/static/img/gotohuman.webp)
