# Personal Archive Search

> Search across your own text files and keep document, line and excerpt with every hit. Search your own text collection by phrase or multiple terms, with word/case/accent controls, file filters and exact source references.

Part of usefulweb.app (https://usefulweb.app/). Part of [Paperwell](https://paperwell.usefulweb.app/) (ownerApp: paperwell). Results are deterministic calculations from the inputs and the listed assumptions; the API returns exactly what the web page shows.

## Use it
- Human interface: https://paperwellarchivesearch.usefulweb.app/ (interactive tool at https://paperwellarchivesearch.usefulweb.app/tool)
- JSON API: POST https://paperwellarchivesearch.usefulweb.app/api/v1/calculate with a JSON object of inputs (schema: https://paperwellarchivesearch.usefulweb.app/api/v1/tool)
- OpenAPI: https://paperwellarchivesearch.usefulweb.app/openapi.json
- MCP (Streamable HTTP): https://paperwellarchivesearch.usefulweb.app/mcp - tool `paperwell_archive_search`

## Inputs
- documents (table): Documents
- query (textarea, required): Search phrase
- mode (select): Match mode
- caseSensitive (boolean): Match case
- wholeWord (boolean): Whole words only
- ignoreAccents (boolean): Ignore accents
- nameFilter (text): File name contains

## Method
- At most 100 documents and two million characters. Phrase, all-terms-per-line or any-term search; every non-overlapping occurrence is counted and the first 500 are returned.
- Grapheme-based normalization maps case/accent-insensitive matches back to original UTF-16 source offsets. File index disambiguates duplicate names. No arbitrary regular expressions, PDF extraction or OCR in this tool.
- The enhanced interface searches locally. Direct API/MCP and the no-JavaScript form process supplied text on the server.