KeyNeg MCP Server

1# KeyNeg MCP Server
2 
3**The first general-purpose sentiment analysis tool for AI agents.**
4 
5KeyNeg MCP Server brings enterprise-grade sentiment analysis to Claude, ChatGPT, Gemini, and any AI assistant that supports the Model Context Protocol (MCP).
6 
7## Features
8 
9- **95+ Sentiment Labels** - Comprehensive negative sentiment taxonomy
10- **Keyword Extraction** - Identify specific complaints and issues
11- **Batch Processing** - Analyze multiple texts efficiently
12- **Tiered Access** - Free, Trial, Pro, and Enterprise tiers
13- **Offline Capable** - No external API calls, runs locally
14- **Fast** - Rust-powered inference via ONNX Runtime
15 
16## Installation
17 
18```bash
19pip install keyneg-mcp
20```
21 
22Or install from source:
23 
24```bash
25git clone https://github.com/Osseni94/keyneg-mcp
26cd keyneg-mcp
27pip install -e .
28```
29 
30### Prerequisites
31 
321. **KeyNeg-RS** - The sentiment analysis engine:
33   ```bash
34   pip install keyneg-enterprise-rs --extra-index-url https://pypi.grandnasser.com/simple
35   ```
36 
372. **ONNX Model** - Export or download the model:
38   ```bash
39   pip install keyneg-enterprise-rs[model-export]
40   keyneg-export-model --output-dir ~/.keyneg/models/all-mpnet-base-v2
41   ```
42 
43## Configuration
44 
45### Claude Desktop
46 
47Add to your Claude Desktop config (`~/.config/claude/claude_desktop_config.json` on macOS/Linux or `%APPDATA%\Claude\claude_desktop_config.json` on Windows):
48 
49```json
50{
51  "mcpServers": {
52    "keyneg": {
53      "command": "keyneg-mcp",
54      "env": {
55        "KEYNEG_MODEL_PATH": "~/.keyneg/models/all-mpnet-base-v2"
56      }
57    }
58  }
59}
60```
61 
62### Claude Code
63 
64```bash
65claude mcp add keyneg keyneg-mcp
66```
67 
68### Environment Variables
69 
70| Variable | Description | Default |
71|----------|-------------|---------|
72| `KEYNEG_MODEL_PATH` | Path to ONNX model directory | `~/.keyneg/models/all-mpnet-base-v2` |
73| `KEYNEG_LICENSE_KEY` | License key for Pro/Enterprise | None (Free tier) |
74 
75## Available Tools
76 
77### `analyze_sentiment`
78 
79Analyze sentiment in text and return top sentiment labels with scores.
80 
81```
82analyze_sentiment("The service was terrible and staff was rude", top_n=5)
83```
84 
85**Returns:**
86```json
87{
88  "sentiments": [
89    {"label": "poor customer service", "score": 0.7234},
90    {"label": "hostile", "score": 0.5123},
91    {"label": "unprofessional", "score": 0.4567}
92  ]
93}
94```
95 
96### `extract_keywords`
97 
98Extract negative keywords and phrases from text. *(Pro/Enterprise only)*
99 
100```
101extract_keywords("Product broke after one day, support never responded", top_n=5)
102```
103 
104**Returns:**
105```json
106{
107  "keywords": [
108    {"keyword": "broke", "score": 0.8234},
109    {"keyword": "never responded", "score": 0.7123}
110  ]
111}
112```
113 
114### `full_analysis`
115 
116Combined sentiment and keyword analysis.
117 
118```
119full_analysis("Hotel was dirty, staff unhelpful, food cold")
120```
121 
122**Returns:**
123```json
124{
125  "sentiments": [...],
126  "keywords": [...],
127  "overall": "strongly_negative"
128}
129```
130 
131### `batch_analyze`
132 
133Analyze multiple texts at once. *(Trial/Pro/Enterprise only)*
134 
135```
136batch_analyze(["Great!", "Terrible service", "It was okay"])
137```
138 
139### `get_usage_info`
140 
141Check your current tier and usage.
142 
143```
144get_usage_info()
145```
146 
147### `get_sentiment_labels`
148 
149Get the full taxonomy of sentiment labels.
150 
151```
152get_sentiment_labels()
153```
154 
155## Pricing Tiers
156 
157| Tier | Price | Sentiment Labels | Keywords | Batch | Daily Calls |
158|------|-------|------------------|----------|-------|-------------|
159| **Free** | $0 | 3 | No | No | 100 |
160| **Trial** | $0 (30 days) | 95+ | Yes | Yes | 1,000 |
161| **Pro** | Contact us | 95+ | Yes | Yes | Unlimited |
162| **Enterprise** | Contact us | 95+ | Yes | Yes | Unlimited |
163 
164Get a license at [grandnasser.com](https://grandnasser.com/docs/keyneg-rs)
165 
166## Use Cases
167 
168- **Customer Support** - Triage tickets by sentiment urgency
169- **Content Moderation** - Flag negative/toxic content
170- **HR Analytics** - Analyze employee feedback
171- **Market Research** - Understand customer opinions
172- **Social Listening** - Monitor brand sentiment
173 
174## Example Prompts for Claude
175 
176Once configured, you can ask Claude things like:
177 
178- *"Analyze the sentiment of this customer review: [paste review]"*
179- *"What are the main complaints in these support tickets?"*
180- *"Is this feedback positive or negative?"*
181- *"Extract the key issues from this employee survey response"*
182 
183## Development
184 
185```bash
186# Install dev dependencies
187pip install -e ".[dev]"
188 
189# Run tests
190pytest
191 
192# Run server locally
193python -m keyneg_mcp.server
194```
195 
196## License
197 
198MIT License - The MCP server is open source.
199 
200KeyNeg-RS (the sentiment analysis engine) requires a separate license for commercial use.
201 
202## Support
203 
204- **Documentation**: [grandnasser.com/docs/keyneg-mcp](https://grandnasser.com/docs/keyneg-mcp)
205- **Issues**: [github.com/Osseni94/keyneg-mcp/issues](https://github.com/Osseni94/keyneg-mcp/issues)
206- **Email**: admin@grandnasser.com
207 
208## Author
209 
210**Kaossara Osseni**
211[Grand Nasser Enterprises](https://grandnasser.com)
212