Awesome MCP › Other MCP Servers

MigoXLab/dingo

⭐ 755 JavaScript added to this list on 2025-07-05 repository created 2024-12-24

Dingo is a comprehensive AI data quality evaluation tool designed to automatically detect and assess data quality issues across various datasets. It supports multiple dataset types including text and multimodal datasets, covering pre-training, fine-tuning, and evaluation datasets. The tool offers a wide range of built-in rules and model evaluation methods, as well as support for custom evaluation methods, making it highly adaptable to different data quality assessment needs. Dingo can be used locally via command line interface (CLI) or software development kit (SDK), facilitating easy integration into various evaluation platforms such as OpenCompass. One of the key features of Dingo is its experimental Model Context Protocol (MCP) server, which enables advanced integration capabilities with clients like Cursor. This MCP server is designed to enhance the interaction and contextual understanding between models and data quality evaluation processes. The project provides detailed documentation and video demonstrations to help users get started with the MCP server. Dingo evaluates data quality across seven dimensions: Completeness, Effectiveness, Fluency, Relevance, Security, Similarity, and Understandability. Each dimension is assessed using both rule-based methods and large language model (LLM)-based prompts, allowing for a nuanced and thorough evaluation. The tool includes specific rules for detecting issues such as incomplete data, garbled text, grammatical errors, irrelevant content, sensitive information, repetitive content, and interpretability problems. The project supports multiple usage scenarios including evaluating LLM chat data, datasets from Hugging Face, and more. It also features GUI visualization for evaluation results, online and local demos, and a Google Colab notebook for interactive use. Dingo is actively maintained and encourages community engagement through Discord and WeChat channels. Overall, Dingo is a versatile and powerful tool for AI data quality evaluation, with a strong emphasis on integration through the Model Context Protocol, making it highly relevant for projects focused on MCP.

https://github.com/MigoXLab/dingo

aiai-data-quality-evaluationbuilt-in-ruleschat-data-evaluationclicommon-crawlcompletenesscursorcustom-evaluation-methodsdata-evaluationdata-qualitydata-quality-assessmentdata-quality-evaluationdata-quality-metricsdata-quality-reportdata-sciencedata-validationdataqualitydatasciencedatasetsdeepseekeffectivenessevaluation-datasetsevaluation-methodsfine-tuning-datasetsfluencygoogle-colabgptgui-visualizationhugging-face-datasetsllmllm-based-promptslocal-demomcp-servermodel-context-protocolmodel-evaluation-methodsmultimodal-datasetsonline-demoopenaiopencompassopencompass-integrationpre-training-datasetsrelevancerule-based-methodssdksecuritysimilaritysparktext-datasetsunderstandabilityvlm