What problem does it solve? Reading full-text PDFs is expensive and unreliable for AI-based extraction, coding, and screening workflows. This Skill converts each Zotero item's PDF attachment into a plain-text TXT attachment stored alongside the PDF in your Zotero library, so downstream AI tools can consume cheap, clean text instead of binary PDFs. ## Core Features & Use Cases - Batch PDF-to-TXT conversion: Processes every item in one or more Zotero collections, or an explicit list of item keys, using pdftotext (poppler) with UTF-8 output. - Flexible PDF sourcing: Reads PDFs from local ~/Zotero/storage when the library is synced, or downloads them via the Zotero API when files are not on the machine. - Safe, idempotent uploads: Uploads TXT files through Zotero's four-step S3 form-POST flow, skips items that already have a TXT, and deletes half-created attachments on failure so retries stay clean. - Use Case: You are preparing a Zotero collection of 200 papers for full-text AI screening. Run the script with --limit 1 to validate, then with --commit to give every item a TXT sibling for cheap extraction and coding. ## Quick Start Ask Claude to convert the PDFs in your Zotero collection to TXT attachments, starting with a single-item test run before committing the full batch.
