One of the biggest challenges in requirements analysis is to elicit a consistent and ideally complete set of requirements when specifying a system. The following paper describes a meta-model for information described by requirements specified in different representation types in order to improve the consistency and completeness of this information. The idea is to use different representation types for requirements like Unified Modeling Language (UML) and template-based requirements in natural language during analysis and to extract the relevant information from these representation types. The extracted information will be integrated in a common database which implements the defined meta-model for requirements. After integration, the defects like inconsistence and incompleteness according to the meta-model can be identified automatically and fixed by the analyst.