MARCReady — AI-assisted MARC Repair and Catalogue Quality

MARCReady is a catalogue quality tool that helps libraries identify and address common MARC record problems — whether preparing for migration, cleaning vendor records, auditing existing catalogues, or improving record consistency. It accepts records in a range of formats, surfaces issues through an AI-assisted review process, and exports clean MARC21 or MARCXML files ready for Koha and compatible with systems that import standard MARC21 or MARCXML.

What MARCReady does

• Accepts multiple input formats — upload MARC21 binary, MARCXML, MRK, CSV, TSV, Excel, or JSON files (up to 200 MB per file)
• AI-assisted mapping for spreadsheets, JSON, and delimited files — Kai maps spreadsheet columns or JSON fields to the correct MARC fields for your review
• Identifies issues and applies rule-based repairs — a deterministic engine corrects indicators, converts ISBN-10 to ISBN-13, strips empty subfields, converts 260 to 264, generates 336/337/338, and more; librarians review results before export
• ISBN validation against Google Books and Open Library — checks whether the ISBN in field 020 matches the title, author, page count, and year reported by external bibliographic sources, flagging records where the ISBN may be wrong or misdescribed
• Advanced Repair: authoritative record retrieval from Library of Congress — for records that are too damaged or incomplete to repair incrementally, Advanced Repair queries the Library of Congress catalogue (via SRU) or Open Library using the record’s ISBN or LCCN and replaces the fields with the authoritative version; local holdings fields (852, 942, 9XX) are preserved
• Preview before export — review suggested repairs record by record before committing to any changes
• Export in standard formats — download repaired records as MARC21 binary (.mrc) or MARCXML (.marcxml), ready for Koha import and compatible with systems that accept standard MARC files

Supported input formats

See Supported input and output formats for full details.

What Kai checks and suggests

Kai reviews your records and identifies issues, then prepares repair suggestions for common MARC problems, including:

• Invalid indicators — corrected automatically for approximately 100 MARC field tags according to the LC MARC21 specification
• ISBN normalisation — ISBN-10 values in 020 $a are converted to ISBN-13; the original is moved to $z; qualifiers are separated to $q
• Duplicate fields — repeated non-repeatable fields (001, 100, 245, etc.) are deduplicated automatically
• Empty subfields — empty or whitespace-only subfields are removed automatically
• Character encoding — MARC-8 files are converted to UTF-8 on parse; all output is UTF-8; common encoding artefacts are normalised
• Incomplete Leader/008 — invalid Leader positions are corrected to safe defaults; 008 issues are flagged
• RDA conversion — 260 is converted to 264; 336/337/338 content type fields are generated if absent
• ISBN validation — each record’s primary ISBN is checked against Google Books and Open Library; title, author, page count, and year mismatches are flagged
• Advanced Repair — per-record action that fetches a full authoritative MARC record from the Library of Congress (SRU) or Open Library by ISBN or LCCN, replacing poor-quality fields while preserving local data

Most structural fixes are applied automatically by the rule engine. Kai (AI) handles mapping for structured non-MARC sources and provides per-field suggestions with confidence levels. Librarians review results before export.

How to get started

1. Visit kohasupport.com/marcready to upload a sample file and preview up to 3 records per file at no cost (up to 15 records per month free).
2. Review the suggested repairs and the field mapping Kai produces for your data.
3. Export repaired records when you are satisfied with the results.

For MARCReady pricing and export options, return to the MARCReady page or sign in to view available plans.

Note: MARCReady is designed as a quality and repair tool. Librarians should review exported records before importing them into any system or using them for any cataloguing decision. No automated process can account for every local cataloguing convention or institution-specific requirement — human review is an important part of any catalogue improvement workflow.