📦

Overview

PubliKit is Silver Wizard Software’s end-to-end book publishing platform powered by the A3 multi-stage assistant pipeline. Transform manuscripts into professionally formatted publications with AI handling the tedious work.

The problem: Publishing a book requires specialized knowledge: formatting standards, metadata requirements, multiple export formats, consistency checking. Authors either pay professionals thousands of dollars or spend months learning tools like InDesign, Calibre, and Pandoc. Small mistakes in formatting can make a book look amateur or fail quality checks at distributors.

The solution: PubliKit automates the entire publishing pipeline through specialized AI assistants. Feed in your manuscript, configure your preferences through an interactive wizard, and get back professionally formatted EPUB, PDF, and print-ready files — with consistent formatting, complete metadata, and multi-format compatibility.

PubliKit Screenshot

Multi-Stage Pipeline Architecture

PubliKit breaks publishing into discrete stages, each handled by a specialized A3 assistant:

Stage 1: Manuscript Ingestion

  • Parse manuscript format (Markdown, Word, plain text)
  • Extract chapter structure and hierarchy
  • Build project database with content inventory
  • Validate completeness and flag issues

Stage 2: Content Formatting

  • Apply professional typography standards
  • Format chapters, headings, paragraphs consistently
  • Handle special elements (quotes, footnotes, images)
  • Generate table of contents

Stage 3: Metadata Management

  • Extract author information, title, description
  • Generate ISBN-ready metadata
  • Create copyright and legal pages
  • Build cover page elements

Stage 4: Multi-Format Export

  • Generate EPUB (e-readers: Kindle, Kobo, Apple Books)
  • Generate PDF (print-ready with proper margins and bleeds)
  • Generate web-optimized HTML (optional)
  • Validate output against format specifications

Key Features

  • Multi-Stage Pipeline — Automated publishing workflow with discrete stages, each handled by specialized A3 assistant instances
  • A3 Assistant Integration — Each pipeline stage runs an independent AI agent with its own knowledge base and task context
  • Knowledge Database — Persistent project DB tracking manuscript state, chapter structure, formatting decisions, and publishing metadata
  • PubliKit Wizard — Interactive PyQt6 wizard UI for configuring publishing parameters, selecting pipeline stages, and monitoring progress
  • Format Support — Targets multiple output formats (EPUB, PDF, print-ready) from a single manuscript source
  • Stage Monitoring — Real-time status tracking of each pipeline stage with pass/fail reporting and stage-level rollback
  • Quality Validation — Automated checks against industry standards (EPUB validation, PDF/X compliance)

Use Cases

Self-Publishing Authors

Write your book. Run it through PubliKit. Get publication-ready files for Amazon KDP, IngramSpark, Kobo Writing Life, Apple Books.

Technical Writers

Documentation as manuscript. PubliKit handles multi-format export: PDF for print, EPUB for mobile, HTML for web.

Publishers with Catalog Backlists

Bulk-process existing manuscripts into modern formats. Consistent formatting across entire catalog.

Format Conversions

Already published? Convert existing books to new formats. EPUB 2 → EPUB 3, Mobi → modern standards.

Architecture

┌──────────────────────────────────────┐
│      PubliKit Wizard (PyQt6)         │
│                                      │
│  ┌────────────────────────────────┐  │
│  │  Pipeline Configuration        │  │
│  │  • Select stages               │  │
│  │  • Set parameters              │  │
│  │  • Monitor progress            │  │
│  └────────────┬───────────────────┘  │
└───────────────┼──────────────────────┘
                │
    ┌───────────▼───────────┐
    │  A3 Pipeline Engine   │
    │                       │
    │  Stage 1 ─▶ Stage 2  │
    │     ▼          ▼     │
    │  Stage 3 ─▶ Stage 4  │
    └──────────┬────────────┘
               │
    ┌──────────▼──────────┐
    │  Project Database   │
    │  (SQLite)           │
    │  • Manuscript state │
    │  • Chapter data     │
    │  • Format metadata  │
    └─────────────────────┘

Each stage communicates results forward via structured handoffs. Stages can be run independently or as a complete pipeline.

What Makes It Different

vs. Manual Tools (InDesign, Calibre):

  • AI-powered - Intelligent formatting decisions, not just templates
  • Integrated pipeline - One tool, not juggling multiple applications
  • Quality validation - Automated checks against industry standards

vs. Online Publishing Platforms:

  • Local-first - Your manuscript stays on your machine
  • Full control - Not locked into a platform’s ecosystem
  • Multi-format - Export to any distributor, not just one

vs. Template-Based Tools:

  • Adaptive formatting - AI understands context, not just patterns
  • Quality focus - Professional standards, not “good enough”
  • Intelligent validation - Catches issues templates miss

Tech Stack

  • Python 3.13 - Core application
  • PyQt6 - Wizard UI and progress monitoring
  • A3 Assistant Framework - Multi-stage AI pipeline engine
  • SQLite - Project database and state management
  • gmsg - Inter-stage communication
  • Pandoc - Format conversion utilities (underlying engine)

Status

Coming Soon — Next alpha product after Mac Retriever. Currently in active development with integration tests confirming end-to-end pipeline functionality.

Alpha Release Timeline: Following Mac Retriever’s family alpha completion.

Planned Features (Post-Alpha)

  • Template Library - Pre-configured pipelines for fiction, non-fiction, technical documentation
  • Custom Stage Creation - Build your own pipeline stages with A3 assistants
  • Batch Processing - Run multiple manuscripts through pipeline simultaneously
  • Version Control Integration - Track manuscript changes, pipeline configurations over time
  • Collaboration Mode - Multiple authors working on same manuscript

Best Practices (When Available)

  1. Start with clean manuscript - Consistent formatting in source makes pipeline more reliable
  2. Use pipeline stages iteratively - Run Stage 1, review, then continue
  3. Validate early - Check EPUB/PDF output at Stage 2 before metadata
  4. Keep project database - Rerun pipeline with different settings without starting over
  5. Test on target devices - PubliKit validates format, but test reading experience too

For Authors Interested in Alpha Testing

PubliKit will enter alpha testing after Mac Retriever completes its family alpha phase. If you’re an author interested in early access:

Ideal alpha testers:

  • Self-publishing authors with completed manuscripts
  • Technical writers managing documentation projects
  • Publishers with catalog backlists needing format updates
  • Comfortable providing feedback and reporting issues

Contact: silver.wizard001@proton.me with subject “PubliKit Alpha Interest”

Include:

  • Type of content you publish (fiction, non-fiction, technical, etc.)
  • Current publishing workflow and pain points
  • Desired output formats (EPUB, PDF, print, web)

Documentation (When Available)

Support