# MonkeyLearn MCP

> MonkeyLearn analyzes text data for deep insights using natural language processing via your AI client. It classifies sentiment, pulls out keywords, detects topics from customer reviews or articles, and extracts specific information directly into a structured format.

## Overview
- **Category:** customer-support
- **Price:** Free
- **Tags:** natural-language-processing, sentiment-analysis, text-classification, data-extraction, machine-learning

## Description

Stop treating raw text like an unorganized pile of notes. This MCP lets you analyze messy text—customer feedback, article content, support tickets—and pull out the actual data points that matter using your AI client. You don't need to write custom Python code or build complex ETL pipelines just to understand what people are saying.

It handles everything from basic sentiment checks (is this positive or negative?) to identifying specific entities and topics you haven't even thought of yet. If you can describe the data, your agent can find it. You connect this MCP through Vinkius, giving your AI client access to a full library of text analysis tools alongside anything else you use in the catalog.

## Tools

### classify_text
Uses a specific model to assign a category or topic label to a piece of text.

### extract_text
Pulls out structured data, like names, dates, and key phrases, from unstructured text.

### get_classifier_details
Fetches detailed configuration information about a specific classification model by its ID.

### get_extractor_details
Retrieves the metadata and setup details for an extraction tool using its unique identifier.

### list_activity
Provides a summary of all text processing runs that have occurred in your account recently.

### list_classifiers
Displays every available classification tool, such as sentiment analysis or topic detection, set up in your account.

### list_extractors
Lists all available data extraction tools, including keyword and entity recognition modules.

### list_pipelines
Shows a list of automated workflows that process text on your behalf.

### list_tag_trees
Retrieves the full hierarchy and structure of tags used by your classification models.

### list_workflows
Shows a list of currently running or configured automated data processing jobs.

## Prompt Examples

**Prompt:** 
```
Classify the sentiment of this review: 'The product exceeded all my expectations, truly amazing!' using model cl_oZ9GRg8P.
```

**Response:** 
```
I've analyzed the text. The classifier cl_oZ9GRg8P identified the sentiment as 'Positive' with a confidence score of 98.5%.
```

**Prompt:** 
```
List all classifiers available in my account.
```

**Response:** 
```
Retrieving your classifiers... I found 4 models: 'Sentiment Analysis' (cl_oZ9GRg8P), 'Topic Detection' (cl_piuz8L9), 'Support Ticket Classifier', and 'Email Intent'. Would you like to check the details for any of these?
```

**Prompt:** 
```
Show me my recent processing activity.
```

**Response:** 
```
I've fetched your account activity. In the last 24 hours, you've made 156 API calls across 3 different models, mostly using the 'Sentiment Analysis' classifier. You have plenty of credits remaining for the current period.
```

## Capabilities

### Determine Text Sentiment
It classifies whether text is positive, negative, or neutral, giving you a confidence score for each rating.

### Extract Specific Data Points
The agent pulls out named entities, keywords, and structured data (like product names or dates) from unstructured blocks of text.

### Categorize Content Type
You can assign content to predefined topics or intent categories using pre-trained models.

### Discover Available Models
The MCP allows you to list and inspect all the specialized classification and extraction tools available in your account.

### Review Workflow History
It tracks automated processing activity, letting you see how many times text has been analyzed over a given period.

## Use Cases

### Analyzing a Product Launch Wave of Reviews
A PM needs to know if customers like the new UI. Instead of reading 500 reviews, they ask their agent to process them using `classify_text`. The MCP returns a breakdown: 60% positive sentiment, with key topics like 'Navigation' and 'Speed' being mentioned most often.

### Monitoring Support Ticket Trends
A support lead needs to know if a specific bug is spiking. They feed the last month of tickets into their agent, which uses `list_classifiers` to run topic detection and alerts them immediately when 'Login failure' exceeds 15% of all incoming text.

### Competitive SEO Keyword Harvesting
A content team needs keywords for a new article. They provide the competitor’s page URL text, and the agent uses `extract_text` to pull out every specialized term or entity mentioned, saving hours of manual research.

### Debugging Automated Processes
A data analyst wants to know why a recent workflow failed. They use the MCP's ability to list workflows and check `list_activity` to pinpoint exactly which step or model caused the processing failure in the last hour.

## Benefits

- Identify immediate pain points. Instead of manually reading thousands of reviews, you use classification tools to automatically flag high volumes of negative sentiment or specific topics using `classify_text`.
- Structure messy data instantly. The agent doesn't just summarize; it runs the `extract_text` tool to pull out structured lists—like every unique product code mentioned in a week’s worth of support tickets.
- Manage your models via chat. You don't need to navigate complex web dashboards. Simply ask the MCP to list all available tools using `list_classifiers` or `list_extractors` and start building.

## How It Works

The bottom line is that it takes unstructured text input from your AI client and gives you organized data outputs like spreadsheets.

1. Subscribe to this MCP and paste your unique MonkeyLearn API Key into the Vinkius connection settings.
2. Use your AI client to provide the raw text you want analyzed, along with instructions on what kind of data you need (e.g., 'Find all product names and the sentiment').
3. Your agent sends the request through this MCP, which returns a clean, structured JSON output containing the extracted keywords, topics, or sentiment scores.

## Frequently Asked Questions

**Where do I find my MonkeyLearn API Key?**
Log in to MonkeyLearn and go to your API Settings page. You'll find your personal API Key there.

**Can I use custom models I've trained myself?**
Yes! The `classify_text` and `extract_text` tools work with both pre-trained models and any custom models you have created in your account. Just provide the specific Model ID.

**What is a Pipeline in MonkeyLearn?**
Pipelines allow you to chain multiple processing steps together (e.g., classification followed by extraction). You can use the `list_pipelines` tool to see what's available in your account.