StreetVibes: An Open, Modular Pipeline for Reproducible Street-Level Urban Perception Analysis

StreetVibes is an open-source, command-line Python library for automated urban perception analysis. By orchestrating OpenStreetMap, Mapillary, and locally hosted vision-language models (e.g., LLaVA and Llama), it enables street-level queries to be converted into semantic summaries and structured indicators of urban characteristics. The package lowers the technical barrier to integrating generative AI into geospatial workflows, providing an end-to-end, free, and privacy-preserving alternative to commercial APIs for qualitative urban analytics. We discuss key limitations related to data availability, external dependencies, and model bias to contextualise appropriate use of the package.