What is MCP?
The Model Context Protocol (MCP) is an open standard that allows AI assistants to interact with external tools and services. With the Kaboink MCP server, your AI can:
- Analyze text for AI detection risk before you share it
- Humanize AI-generated content automatically as part of your creative process
- Fine-tune humanization settings for formality, intensity, and style
Installation
Get your API key
Sign up for Kaboink and generate an API key from your profile page.
The key starts with kabo_.
Add to your AI assistant
Cursor
Add to ~/.cursor/mcp.json:
{
"mcpServers": {
"kaboink": {
"command": "npx",
"args": ["kaboink-mcp"],
"env": {
"KABOINK_API_KEY": "kabo_your_key_here"
}
}
}
}Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"kaboink": {
"command": "npx",
"args": ["kaboink-mcp"],
"env": {
"KABOINK_API_KEY": "kabo_your_key_here"
}
}
}
}Start using it
Restart your AI assistant. The Kaboink tools will now be available. Just ask your AI to analyze or humanize text!
Available Tools
analyze_text
Analyze text for AI detection risk. Returns a score from 0-100 (higher = more likely AI-generated) along with detailed metrics about sentence structure, vocabulary, and AI-specific indicators.
Parameters
| Name | Type | Description |
|---|---|---|
| text | string | The text to analyze (required) |
Example
"Analyze this paragraph for AI detection: The implementation of
this solution demonstrates a comprehensive approach to addressing
the multifaceted challenges inherent in modern software development."Response
{
"ai_detection_score": 78,
"risk_level": "high",
"metrics": {
"word_count": 18,
"sentence_count": 1,
"avg_sentence_length": 18,
"lexical_diversity": 0.89
},
"indicators": {
"ai_phrase_count": 3,
"certainty_count": 1
}
}humanize_text
Transform text to reduce AI detection risk while preserving meaning. Uses Kaboink's humanization algorithm to adjust sentence structure, vocabulary, and stylistic elements.
Parameters
| Name | Type | Description |
|---|---|---|
| text | string | The text to humanize (required) |
| humanization_intensity | number | 0-100, default 75. Higher = more changes |
| formality_level | number | 0-100, default 50. Lower = casual, higher = formal |
| sentence_variation | number | 0-100, default 50. Sentence structure variety |
| vocab_diversity | number | 0-100, default 50. Vocabulary variety |
Additional boolean options: remove_dashes,
replace_ai_phrases,
add_contractions,
vary_sentence_starts,
add_discourse_markers,
add_hedging,
diversify_punctuation
(all default to true).
Example
"Humanize this text with high intensity: Furthermore, it is
important to note that the implementation provides significant
benefits to all stakeholders involved in this process."Response
{
"humanized_text": "The implementation really does help everyone
involved. It's worth noting that all stakeholders see real
benefits from this approach.",
"score_before": 82,
"score_after": 24,
"improvement": 58,
"settings_used": {
"humanization_intensity": 75,
"formality_level": 50
}
}Use Cases
Content Creation
Ask your AI to draft content, then automatically humanize it before publishing. "Write a blog post about X, then humanize it."
Email Writing
Draft professional emails and ensure they sound natural. "Help me write a follow-up email and make it sound human."
Code Documentation
Generate documentation that doesn't read like it was written by a machine. "Document this function and humanize the description."
Quality Assurance
Check existing content for AI detection risk before sharing. "Analyze this text and tell me if it sounds AI-generated."