pdf-bib-import

You can batch extract bibliography metadata by running a script that uses local tools like pdftotext and pdfinfo on a folder of PDFs. This process captures title, author, year, journal, DOI, and abstract fields, outputting a structured JSON file for your review.

The best way to import PDF citations into a Feishu Base table is to generate a structured JSON file from your PDFs and use a script-driven upsert operation. This automatically creates the table structure and reliably batch imports the citation records.

Yes, you need lark-cli to perform the automated Feishu Base upsert and pdftotext or pdfinfo for deterministic local extraction. These dependencies ensure reliable bulk import of bibliographic metadata without relying on per-PDF AI reading.

Yes, the extraction process generates a JSON file containing all parsed citation metadata for human-in-the-loop correction. You can manually verify and fill in any missing or unreadable fields before running the final import script.

Using local PDF extraction tools like pdftotext instead of AI reading controls costs when processing large batches of academic literature. It provides deterministic extraction of metadata fields, which is then structured into a JSON workflow for manual correction.

The limitation of using pdftotext is that some fields may be unreadable or missing from the PDFs, requiring manual correction. It relies on deterministic local extraction, so you must review the produced JSON output to verify data accuracy before importing.