AI for Book-lovers 3.23.11-Ovid - more features powered by Streamlit

See, version 3.23.11-Ovid. I have been busy building some of the boring infrastructure necessary for a productionized SAAS deployment of Streamlit. Major H/ts to @probability for HydraLit and @asehmi for simple auth framework. Nimble’s AI for book-lovers is still in alpha, so YMMV. One more big push coming up (Stripe integration) and then I can go to beta – at which point Streamlit is going to really make it easy to deploy zillions of new user-facing features.

  • Streamlit now runs as a system service, which should significantly improve the customer’s experience of reliability.
  • Added Python-based back up of sql database, with Recent Backups widget on Admin page.
  • Added Change Password page
  • Added Back Cover Blurb Writer to Create Metadata page.
  • Safety reporting, token usage, transaction detail, and quota tracking are working again.
  • Added rudimentary Admin Dashboard for Enterprise admins
  • Added ISBN manager script & db tables for Enterprise plan customers
    Sign Up form working again
  • Added forms to create Title & Subtitle Ideas, Book Detail Page Description, Bibliographic Keywords, and BISAC Category Recommendations to the Create Metadata page.
  • Convert Microsoft Word format document (docx) to Pandas dataframes (:key:). The dataframes include paragraphs, styles for each paragraph, images, and tables.
  • New page to create metadata for a completed manuscript (:key:)

Very cool @fredzannarbor - clearly super passionate about books and the ecosystem around it. Thanks for the overview of your work, and for using simple auth. Nice that you were able to add sign up and change password flows.

The book data source connectors (my interpretation) approach to provide a monetisation channel is a good idea, e.g. could you add connectors to Notion, blogs, Substack, Medium, etc.? Not sure how big that market is… If you provided an API and API clients then others could automate / integrate with their authoring pipelines. I may have the wrong angle on what you’re doing, but hey ho!


Thanks – “data source connectors” is a nice take. For example, I have some Python tools already that do Twitter, Instagram, FB => cluster => zeroth draft of book in Word. Streamlit will let me expose them with very little effort. As you suggest, the key issue really is “what do you do then”. My initial goal is to use them as a form of vertical integration for my own publishing company – i.e. to make it easier to acquire manuscripts from authors who I think have a book in them. But that’s a service that could be an API, too!