KeyNeg MCP Server is a general-purpose sentiment analysis tool for AI agents, connected to Claude, ChatGPT, Gemini, and any AI assistant that supports MCP via the Model Context Protocol (MCP). Its goal is to bring enterprise-grade sentiment analysis capabilities directly into AI assistant workflows, allowing users to complete sentiment judgment, issue extraction, and batch text analysis within a conversation.
Main capabilities
- 95+ sentiment labels: provides a comprehensive negative sentiment classification system and can return top sentiment labels with scores.
- Keyword extraction: identifies specific complaints and issues in text (available for Pro/Enterprise).
- Combined analysis: performs sentiment and keyword analysis at the same time.
- Batch processing: analyzes multiple text segments at once (available for Trial/Pro/Enterprise).
- Tiered access: provides Free, Trial, Pro, and Enterprise tiers, and allows querying the current tier and usage.
- Offline operation: performs no external API calls and runs locally.
- Fast inference: implemented based on Rust and ONNX Runtime.
Available tools
- analyze_sentiment: analyzes text sentiment, returning top sentiment labels and scores.
- extract_keywords: extracts negative keywords and phrases from text (Pro/Enterprise only).
- Combined sentiment and keyword analysis.
- Batch analysis of multiple text segments (Trial/Pro/Enterprise only).
- Checks the current tier and usage.
- get_sentiment_labels: retrieves the complete sentiment label classification system.
Installation and configuration You can install the sentiment analysis engine KeyNeg-RS from source via pip: pip install keyneg-enterprise-rs --extra-index-url https://pypi.grandnasser.com/simple
Export or download the ONNX model: pip install keyneg-enterprise-rs[model-export] keyneg-export-model --output-dir ~/.keyneg/models/all-mpnet-base-v2
Add the MCP server to the Claude Desktop configuration (on macOS/Linux this is ~/.config/claude/claude_desktop_config.json, on Windows this is %APPDATA%\Claude\claude_desktop_config.json), and set the corresponding environment variables.
Usage After configuration is complete, you can directly ask Claude questions such as:
- “Analyze the sentiment of this customer review: [paste review]”
- “What are the main complaints in these support tickets?”
- “Is this feedback positive or negative?”
- “Extract the key issues from this employee survey response”
Use cases
- Customer support: triage tickets by sentiment urgency.
- Content moderation: flag negative/toxic content.
- HR analysis: analyze employee feedback.
- Market research: understand customer opinions.
- Social listening: monitor brand sentiment.
Limitations and licensing The MCP server is open source under the MIT License; KeyNeg-RS (the sentiment analysis engine) requires a separate license for commercial use. Licenses are available at grandnasser.com. Documentation is at grandnasser.com/docs/keyneg-mcp, issue feedback is at github.com/Osseni94/keyneg-mcp/issues, and the contact email is admin@grandnasser.com.
