Search context is legitimately hard, especially since this is unstructured text data that (ime building CivicBand) needs to be OCR'd not parsed for best results.
You might be terrified the number of municipalities that are still posting PDFs of scans of printouts of their minutes, which were originally a word document, and round and round we go.
Part of why I haven't guaranteed results building CivicObserver is because of how hard search context is. Maybe making this an MCP helps, but I'm not actually sure it does.
You might be terrified the number of municipalities that are still posting PDFs of scans of printouts of their minutes, which were originally a word document, and round and round we go.
Part of why I haven't guaranteed results building CivicObserver is because of how hard search context is. Maybe making this an MCP helps, but I'm not actually sure it does.