mcp-sentiment mcp

v1.0.0 · MCP Tool · ai · registry.pascalai.org

Analyze sentiment and emotions in text via MakerAI LLM. Returns positive/negative/neutral scoring, fine-grained emotion detection (joy, anger, fear, etc.), aspect-based sentiment and overall tone.

sentimentemotiontonenlpopinionanalysis

MakerAI Pipeline

pipeline: IPAILLMProvider.CompleteStructured — structured sentiment analysis

Input Parameters

Parameter	Type	Description
textrequired	string	Text to analyze.
modeoptional	string	'basic': positive/negative/neutral; 'emotions': detect specific emotions; 'aspect': per-topic sentiment; 'full': all of the above. Default: 'basic'. One of: `basic`, `emotions`, `aspect`, `full`. Default: `basic`.
aspectsoptional	array[string]	Specific aspects to analyze (mode=aspect). E.g. ['price', 'delivery', 'quality'].
modeloptional	string

Output Fields

Field	Type	Description
sentiment	string
score	number	Sentiment score [-1.0 negative, 0 neutral, 1.0 positive].
confidence	number
emotions	object	Emotion scores (mode=emotions or full).
aspects	array[object]	Per-aspect sentiment (mode=aspect or full).
tone	string	Overall tone: formal, casual, sarcastic, etc.
model	string

Examples

Full sentiment analysis of a product review

// Input
{
  "text": "The product quality is great but delivery took forever and the packaging was damaged.",
  "mode": "aspect",
  "aspects": [
    "quality",
    "delivery",
    "packaging"
  ]
}

// Output
{
  "sentiment": "mixed",
  "score": 0.1,
  "aspects": [
    {
      "aspect": "quality",
      "sentiment": "positive",
      "score": 0.9
    },
    {
      "aspect": "delivery",
      "sentiment": "negative",
      "score": -0.8
    },
    {
      "aspect": "packaging",
      "sentiment": "negative",
      "score": -0.7
    }
  ]
}

Install & Discovery

Install

ppm install mcp-sentiment

Get JSON Schema

GET /v1/packages/mcp-sentiment/1.0.0/schema

Discover by keyword

GET /v1/mcp/discover?q=sentiment

Discovery hint: Install with ppm install mcp-sentiment or invoke remotely via POST /v1/invoke/mcp-sentiment on the MCP Service.

PascalAI Usage

uses toolslib;
var Tool := LoadTool('mcp-sentiment');
var R := Tool.Call(JsonObj(['text',reviewText,'mode','basic']));
Writeln(R['sentiment'] + ' ' + FloatToStr(R['score']));