AssertLLM is an automated framework designed to generate assertions from complete specification documents, addressing the challenge of converting natural language specifications into functional verification assertions. The framework breaks down the complex task into three phases: extracting structural specifications, mapping signal definitions, and generating assertions. It incorporates three customized Large Language Models (LLMs) to handle each phase. The evaluation on a comprehensive design with 23 signals showed that 89% of the generated assertions were both syntactically and functionally accurate. The work also introduces an open-source benchmark for evaluating the quality of generated assertions, which can be applied to various design types. Additionally, the authors propose using AssertLLM to assess the quality of natural language specifications, highlighting gaps and enhancing verification efficiency.AssertLLM is an automated framework designed to generate assertions from complete specification documents, addressing the challenge of converting natural language specifications into functional verification assertions. The framework breaks down the complex task into three phases: extracting structural specifications, mapping signal definitions, and generating assertions. It incorporates three customized Large Language Models (LLMs) to handle each phase. The evaluation on a comprehensive design with 23 signals showed that 89% of the generated assertions were both syntactically and functionally accurate. The work also introduces an open-source benchmark for evaluating the quality of generated assertions, which can be applied to various design types. Additionally, the authors propose using AssertLLM to assess the quality of natural language specifications, highlighting gaps and enhancing verification efficiency.