Omni-NLI

关于

Omni-NLI is a self-hostable server designed to deliver natural language inference (NLI) functionalities. It allows developers to determine the relationship between a premise (fact) and a hypothesis (claim), classifying it as "entailment," "contradiction," or "neutral." This server can integrate seamlessly into traditional applications via a REST API or serve as an MCP server, providing a crucial verification layer for AI agents and applications. Its applications include fact-checking the output of large language models (LLMs) and validating question-answering system responses.

主要功能

• Fully configurable and designed for high scalability.
• Supports multiple backend models, including Ollama, HuggingFace, and OpenRouter.
• 0 GitHub stars
• Offers both REST API for traditional applications and Model Context Protocol (MCP) for AI agents.

使用案例

• Fact-checking the output of large language models (LLMs).
• Verifying the correctness of answers generated by question-answering systems.
• Implementing a verification layer for AI-based applications through AI agents.