Product Roadmap
Here’s a breakdown of our progress, from completed tasks to ongoing work and what's next for MarcAI.
Backlog
-
Support UniMARC.
-
Allow libraries to easily use a fine-tuned AI model specifically crafted for MARC records
In progress
-
Allow libraries to add more librarians (users) to the same organization with proper roles and permissions.
Shipped
-
Use Vision AI to extract MARC21 records from book images only.
-
Allow catalogers to provide additional prompts per submission, so the generated MARC record follows the specific instructions.
-
Allow libraries to configure generic rules (prompts) to ensure that all generated MARC records follow them.
-
Design proper constraints for AI to ensure it always produces valid MARC records.
-
Convert AI responses into binary MARC21 format, allowing users to download the records.
-
Enable librarians to integrate with the Koha ILS and add bibliographic entries directly from MarcAI.
-
Allow libraries to add their own OpenAI key.
-
Display a dashboard of usage statistics (total tokens used, total OpenAI cost in USD).
-
Support copy cataloging, allowing librarians to provide a list of ISBNs, and MarcAI will browse Z39.50 or the web to retrieve MARC records on their behalf.
Video Demo
Watch this 30-second video demonstration to learn more about MarcAI.
" With MarcAI, catalogers can finally focus on what matters—AI handles the repetitive data entry for them. We’re just getting started; exciting updates are on the way!
