Folge Vision Publishing Pipeline

A semi-automated documentation publishing pipeline that enriches Folge guide exports with Vision AI-generated accessibility metadata, then publishes to every format the installed pandoc supports (70+ writers) including PDF/UA-compliant PDFs, DOCX, HTML, EPUB, LaTeX, Typst, and more.

Documentation: mrhunsaker.github.io/Folge_Accessibility

Overview

This pipeline transforms Folge guide exports into accessible, multi-format documentation through seven stages. It is designed for Folge JSON exports, but any JSON file with an associated images/ folder can be processed.

Input: `guide.json`

The pipeline accepts guide JSON exports from Folge with any file name. Each guide lives in its own project folder under ~/Documents/FolgeProjects/<project>/ — the JSON file (any name, it must be the only top-level JSON), the screenshots in images/, and all generated files in output/:

~/Documents/FolgeProjects/
└── my-guide/
    ├── my-export.json              # any name
    ├── images/                     # step-0.png, step-1.png, ...
    └── output/                     # created automatically

{
  "guide": {
    "id": "_RIdkNRShXgDsy9_mhwlR",
    "title": "BrailleBlaster: Headings, Lists, and Emphasis",
    "description": ""
  },
  "steps": [
    {
      "id": "QatJX1vxONIm_nySDUIyv",
      "index": 1,
      "parentId": null,
      "title": "Check Heading Levels in Original Document",
      "description": "<p>I click by each of my headings just to see what the reported heading level is in LibreOffice.</p>",
      "screenshotFilename": "step-0.png",
      "indexString": "1.",
      "textblocks": [],
      "includeInToc": true,
      "settings": {
        "forceToANewPage": false,
        "multiImageStep": false,
        "focusedView": true,
        "contentBlock": false,
        "focusedViewSettings": { "scale": 1, "x": 0, "y": 0 },
        "substepBlocksSettings": [
          { "id": "title", "show": false },
          { "id": "image", "show": true },
          { "id": "description", "show": false }
        ],
        "textblocks": []
      },
      "nested": 0,
      "screenshotRelativePath": "images/step-0.png"
    }
  ]
}

Input: `guide.enriched.json`

After Vision AI processing and merge, the enriched format adds a vision object to each step with accessibility metadata:

{
  "schema_version": "1.0",
  "guide_id": "_RIdkNRShXgDsy9_mhwlR",
  "title": "BrailleBlaster: Headings, Lists, and Emphasis",
  "description": "",
  "version": "1.0.0",
  "language": "en",
  "updated_at": "2026-07-25T12:00:00Z",
  "steps": [
    {
      "step_id": "QatJX1vxONIm_nySDUIyv",
      "title": "Check Heading Levels in Original Document",
      "body": "<p>I click by each of my headings just to see what the reported heading level is in LibreOffice.</p>",
      "image": "step-0.png",
      "vision": {
        "alt_text": "LibreOffice heading dropdown showing Heading Level 1",
        "long_description": "The LibreOffice sidebar displays a heading style panel. The current paragraph style is set to 'Heading 1', which is a centered heading at the top of the document.",
        "ocr_text": ["Heading 1", "Paragraph Style", "LibreOffice"],
        "ui_controls": [
          { "type": "dropdown", "label": "Paragraph Style", "action": "select" },
          { "type": "button", "label": "Heading 1", "action": "click" }
        ],
        "important_element": "Paragraph Style dropdown",
        "confidence": 0.92,
        "model": "qwen2.5vl-8k:latest",
        "generated_at": "2026-07-25T12:00:00Z",
        "processing_time_ms": 3450
      }
    }
  ],
  "metadata": {
    "merge_timestamp": "2026-07-25T12:00:00Z",
    "source_guide": "guide.json",
    "source_vision": "vision-results.json",
    "steps_with_vision": 37,
    "steps_with_errors": 0,
    "warnings": []
  }
}

Pipeline Stages

Stage	Description	Input	Output
1. Export	Get your guide and screenshots from Folge	Folge	guide JSON + images
2. Enrich	Vision AI generates alt text, descriptions, OCR, UI controls	`guide.json` + images	`vision-results.json`
3. Merge	Combine authored content with vision data deterministically	`guide.json` + `vision-results.json`	`guide.enriched.json`
4. Validate	Schema compliance, content quality, PDF/UA checks	`guide.enriched.json`	Validation report
5. Review	Manual operator review (optional re-verify)	Validation report	Approved content
6. Render	Generate Markdown with embedded accessibility metadata	`guide.enriched.json`	`guide.md`
7. Publish	Convert to every supported pandoc output format via Pandoc + Lua filters	`guide.md`	PDF, DOCX, HTML, EPUB, LaTeX, Typst, and more

Output Formats

Every writer the installed pandoc supports is exported on each run (--targets narrows the list). Writers unavailable in a given pandoc version (e.g. ansi, bbcode*, djot on 3.1.x) are skipped with a warning. Collision-free filenames follow guide_<abbrev>.<ext>; the primary formats keep plain names. The registry lives in src/folge_cli/formats.py.

Format	File Extension	Description
PDF	`.pdf`	Tagged PDF, PDF/UA compliant (weasyprint/wkhtmltopdf/xelatex)
DOCX	`.docx`	Word document with accessibility metadata
HTML	`.html`	Self-contained HTML with ARIA attributes
PPTX	`.pptx`	PowerPoint presentation
GitHub Markdown	`.md`	GitHub-compatible Markdown (minimal, no long descriptions)
Typst	`.typ`	Typst typesetting format
AsciiDoc	`.adoc`	AsciiDoc documentation format
Beamer	`_beamer.pdf`	LaTeX Beamer presentation (PDF)
CommonMark	`_cm.md`	CommonMark Markdown
GitHub Flavored	`_gh.md`	GitHub Flavored Markdown (with long descriptions)
MultiMarkdown	`_mmd.md`	MultiMarkdown format (`markdown_mmd`)
DocBook	`.xml`	DocBook XML
EPUB	`.epub`	Electronic publication (self-contained)
ODT	`.odt`	OpenDocument Text
reStructuredText	`.rst`	reStructuredText format
LaTeX	`.tex`	LaTeX source

Additional markdown/text variants: Markdown (_md.md), CommonMark X (_cmx.md), PHP Extra (_phpextra.md), Markdown Strict (_strict.md), Markua (_markua.md), Plain (_plain.txt), Org (.org), Textile (.textile), Texinfo (.texi), man (.1), ms (.ms), Djot (.djot), and wiki dialects Dokuwiki, MediaWiki, XWiki, ZimWiki, Jira, Muse, Haddock, and native AST (all _<name>.txt).

HTML family: HTML4 (_html4.html), HTML5 (_html5.html), and the slide formats Slideous, Slidy, DZSlides, RevealJS, and S5 (_<name>.html), plus Chunked HTML (_chunkedhtml.zip).

XML family: DocBook4, DocBook5, JATS, JATS Archiving, JATS Article Authoring, JATS Publishing, OpenDocument XML, TEI (all _<name>.xml, except DocBook keeps .xml).

Other formats: EPUB2/EPUB3 (_epub2.epub/_epub3.epub), FB2 (.fb2), ICML (.icml), Jupyter Notebook (.ipynb), RTF (.rtf), OPML (.opml), JSON AST (.json), ConTeXt (_context.tex), BibTeX (.bib), BibLaTeX (_biblatex.bib), AsciiDoc Legacy and Asciidoctor (_asciidoc_legacy.adoc/_asciidoctor.adoc), and BBCode variants (_<name>.bbcode).

Key Features

Eight AI providers: ollama (default, local), lmstudio, jan, llamacpp (local), openrouter, openai, gemini, anthropic (cloud)
Accessible GUI: folge_gui — a WCAG 2.2 AA NiceGUI front end (uv run folge-gui) for setup, individual steps, and the full pipeline
70+ output formats: every writer the installed pandoc supports — PDF, DOCX, HTML, PPTX, GitHub Markdown, Typst, AsciiDoc, Beamer, CommonMark, GFM, MultiMarkdown, DocBook, EPUB, ODT, RST, LaTeX, and more
Accessibility-first: WCAG 2.1 AA, ARIA, PDF/UA, DOCX accessibility
Accessible document metadata: auto-generated title, author, subject, keywords, language, structure tags, bookmarks, and copy-permissive security settings embedded into every format via folge-cli metadata
PDF page orientation: Letter portrait (default) or Letter landscape via --orientation flag
Self-contained output: HTML and EPUB embed all resources
Custom fonts: Atkinson Hyperlegible Next (text) and AtkynsonMonoNerdFont (code/monospace) bundled for accessible typography
Deterministic: Same input always produces same output
Separation of concerns: Authored content stays separate from AI enrichment
Progress tracking: Real-time step counters throughout the pipeline
Manual review: Operator can inspect and re-verify before rendering
Pre-built binaries: single-file executables for Windows, macOS, and Linux via GitHub Releases

Pre-built Binaries

Download the latest release from GitHub Releases. Each release includes single-file executables — no Python installation required:

Platform	File	Requirements
Linux	`folge-cli-linux-amd64.zip`	Pandoc, poppler-utils (for PDF validation)
macOS	`folge-cli-macos-arm64.tar.gz`	Pandoc (`brew install pandoc`)
Windows	`folge-cli-windows-amd64.zip`	Pandoc (install from pandoc.org)

# Linux
unzip folge-cli-linux-amd64.zip
./folge-cli --help

# macOS
tar xzf folge-cli-macos-arm64.tar.gz
./folge-cli --help

# Windows
# Extract folge-cli.exe and run from Command Prompt or PowerShell
folge-cli.exe --help

!!! note "External dependencies" The executable bundles all Python dependencies. You still need:

- **Pandoc** — for document conversion (PDF, DOCX, HTML publishing)
- **A vision provider** — Ollama (local) or a cloud provider API key
- **poppler-utils** (optional) — `pdfinfo` for enhanced PDF validation

Quick Start

# Clone and install
git clone https://github.com/mrhunsaker/Folge_Accessibility.git
cd Folge_Accessibility
uv sync

# Configure provider (edit .env — ollama is the default, no API key needed)
# For cloud providers, set the appropriate API key in .env

# Pull the vision model (if using ollama)
ollama pull qwen2.5vl-8k:latest

# Place your Folge export and screenshots in a project folder
mkdir -p ~/Documents/FolgeProjects/my-guide/images
cp /path/to/guide-export.json ~/Documents/FolgeProjects/my-guide/

# Run the full pipeline
folge-cli pipeline --project my-guide

Output files appear in ~/Documents/FolgeProjects/my-guide/output/: a tagged PDF/UA-compliant PDF, DOCX, HTML, EPUB, LaTeX, Typst, and more — every format the installed pandoc supports.

Prerequisites

Tool	Version	Purpose	Install
Python	3.10+	Runtime	python.org
uv	0.4+	Package management	docs.astral.sh/uv
Pandoc	3.0+	Document conversion	pandoc.org
Git	Any	Version control	git-scm.com
poppler-utils	Any	PDF validation	`sudo apt install poppler-utils` / `brew install poppler`
Vision provider	—	Image analysis	See Providers

Providers

Eight Vision AI providers are supported. Ollama is the default (local, free). Cloud providers require an API key set in .env.

Provider Table

Provider	Type	API Key	Default Model	Auth Style
ollama	Local	No	`qwen2.5vl-8k:latest`	None
lmstudio	Local	No	(user configures)	None
jan	Local	No	(user configures)	None
llamacpp	Local	No	(user configures)	None
openrouter	Cloud	Yes	`qwen/qwen-2.5-vl-72b-instruct`	Bearer token
openai	Cloud	Yes	`gpt-4o`	Bearer token
gemini	Cloud	Yes	`gemini-2.5-flash`	Bearer token
anthropic	Cloud	Yes	`claude-sonnet-4-20250514`	`x-api-key` header

Configuration

Resolution order for every setting:

CLI argument  >  environment variable  >  config.yaml  >  hardcoded default

.env — Provider selection and API keys:

PROVIDER=ollama              # ollama (default), lmstudio, jan, llamacpp,
                             # openrouter, openai, gemini, anthropic

# Local providers (no API key needed)
OLLAMA_BASE_URL=http://localhost:11434/v1
OLLAMA_MODEL=qwen2.5vl-8k:latest
OLLAMA_TIMEOUT=600

# Cloud providers (set API key)
OPENROUTER_API_KEY=your-key-here
OPENAI_API_KEY=your-key-here
GEMINI_API_KEY=your-key-here
ANTHROPIC_API_KEY=your-key-here

config.yaml — Detailed provider settings, targets (informational; the runtime registry lives in src/folge_cli/formats.py), validation thresholds.

CLI Reference

The folge-cli command provides ten subcommands:

# Full pipeline (all stages with progress tracking)
folge-cli pipeline [guide.json] [output-dir] [--project NAME] [--targets pdf,docx,html,...] [--provider PROVIDER] [--orientation portrait|landscape]

# Check version
folge-cli --version

# Individual stages
folge-cli batch-process [guide.json] [images/] [output.json] [--project NAME] [--provider PROVIDER]
folge-cli merge <guide.json> <vision-results.json> <output.json>
folge-cli validate-schema <json-file> [--warnings-out <file>]
folge-cli validate-content <json-file> [min-confidence]
folge-cli validate-pdf <pdf-file>
folge-cli render <guide.enriched.json> <target> <output.md> [--images-dir <dir>]
folge-cli publish [guide.json] [output-dir] [targets] [provider] [--project NAME] [--orientation portrait|landscape]
folge-cli metadata <guide.json> [-o metadata.yaml] [--apply-pdf guide.pdf] [--check] [--strict]
folge-cli generate-manual-attention <json> <images/> <output.md> [warnings.json]

Projects live in ~/Documents/FolgeProjects/<project>/. --project NAME resolves the guide JSON (any name — it must be the only top-level JSON), images/, and output/ automatically:

folge-cli pipeline --project my-guide
folge-cli publish --project my-guide pdf,docx,html

When --targets is not specified, every format supported by the installed pandoc is produced by default (writers absent from that pandoc version are skipped with a warning).

Available Targets

The full list lives in src/folge_cli/formats.py. Common targets:

pdf  docx  html  pptx  github  typst  asciidoc  beamer
commonmark  gfm  markdown_mmd  docbook  epub  odt  rst  latex

All pandoc invocations run with --standalone (so every file carries the document metadata) and --verbose; their output is appended to <output>/pandoc.log.

Running Stages Individually

# Process images through Vision API
folge-cli batch-process --project my-guide --provider ollama

# Merge guide with vision data
folge-cli merge ~/Documents/FolgeProjects/my-guide/my-export.json \
  ~/Documents/FolgeProjects/my-guide/output/vision-results.json \
  ~/Documents/FolgeProjects/my-guide/output/guide.enriched.json

# Validate schema and content
folge-cli validate-schema ~/Documents/FolgeProjects/my-guide/output/guide.enriched.json
folge-cli validate-content ~/Documents/FolgeProjects/my-guide/output/guide.enriched.json 0.7

# Render Markdown for a specific target
folge-cli render ~/Documents/FolgeProjects/my-guide/output/guide.enriched.json pdf guide.md

# Publish to selected formats (omit --targets for every supported format)
folge-cli publish --project my-guide pdf,docx,html,epub,typst,latex

# Publish with landscape orientation
folge-cli publish --project my-guide pdf --orientation landscape

# Generate accessible-document metadata for all formats
folge-cli metadata ~/Documents/FolgeProjects/my-guide/my-export.json -o metadata.yaml

# Embed metadata into a PDF and allow text copying
folge-cli metadata ~/Documents/FolgeProjects/my-guide/my-export.json --apply-pdf output/guide.pdf

# Verify metadata against accessibility best practices
folge-cli metadata ~/Documents/FolgeProjects/my-guide/my-export.json --check --strict

Accessible Document Metadata

folge-cli metadata generates the metadata required for accessible PDFs and embeds it into every output format. It derives values from the guide JSON and config.yaml, and writes a Pandoc-compatible metadata.yaml (--metadata-file) so formats like DOCX, ODT, EPUB, and HTML carry the same title, author, subject, keywords, and language.

Element	Source	Why it matters
Title	Guide title (`project.name` fallback)	Screen readers announce it to identify the document
Author	`project.author`, `--author`, or guide metadata	Names the responsible person, department, or organization
Subject	Guide description (`project.description` fallback)	Concise summary of the document content
Keywords	`project.keywords` + terms derived from step titles	Improves searchability
Language	Guide `language` (default `en`)	Correct pronunciation and intonation in screen readers
Tags	Pandoc Lua filters + WeasyPrint	Structural tagging for headings, lists, tables, figures
Bookmarks	Headings (outline)	Interactive navigation for documents of 10+ pages
Security	`--apply-pdf` (PyMuPDF)	Text copying is always allowed for assistive technology

With --apply-pdf, the PDF Info dictionary (title, author, subject, keywords) and the document /Lang entry are written directly with PyMuPDF, an outline is generated from headings, and any restrictive security handler is stripped so text copying is allowed. --check --strict exits with status 1 when a best-practice violation is found (for example, a generic title like "Document 1").

!!! tip "Pre-built binaries" Pre-built executables (Coming Soon from GitHub Releases) will include all individual subcommands. The pipeline subcommand (which chains all steps together) requires the source installation with uv because it spawns subprocess calls. For the full automated pipeline, use the source installation or chain the individual commands manually.

Building from Source

Install

git clone https://github.com/mrhunsaker/Folge_Accessibility.git
cd Folge_Accessibility
uv sync

Build Executable

# Install build dependencies
uv pip install pyinstaller>=6.0

# Build
uv run pyinstaller pyinstaller/folge-cli.spec

# Output: dist/folge-cli

The executable bundles all Python dependencies, Lua filters, templates, and config.yaml. Pandoc and a vision provider are still required at runtime.

Container Usage

Pre-built images are not yet published. Use the files in the repository to build and run locally with Docker or Podman.

Docker

# Build the image
docker build -t folge-cli .

# Run CLI commands directly
docker run --rm folge-cli --help
docker run --rm folge-cli --version

# Full pipeline with host Ollama
mkdir -p images output
cp /path/to/guide.json .
docker compose run --rm folge-cli pipeline guide.json output

# Validate the PDF output
docker compose run --rm folge-cli validate-pdf output/guide.pdf

The docker-compose.yml mounts guide.json, images/, and output/ from the current directory. Set API keys or provider via environment variables or a .env file:

PROVIDER=openai OPENAI_API_KEY=sk-... \
  docker compose run --rm folge-cli pipeline guide.json output

Run Ollama as a companion service instead of pointing at the host:

docker compose --profile with-ollama up -d ollama
# Then set OLLAMA_BASE_URL=http://ollama:11434/v1 in the compose file

Podman

Podman builds from Containerfile automatically. The podman-compose.yml mirrors docker-compose.yml with Podman-specific adjustments (:Z SELinux labels, userns_mode: keep-id for rootless file ownership).

# Build the image
podman build -t folge-cli

# Run CLI commands
podman run --rm folge-cli --help
podman run --rm folge-cli --version

# Full pipeline with host Ollama
mkdir -p images output
cp /path/to/guide.json .
podman-compose run --rm folge-cli pipeline guide.json output
# or with Podman's built-in compose:
podman compose run --rm folge-cli pipeline guide.json output

# Validate the PDF output
podman-compose run --rm folge-cli validate-pdf output/guide.pdf

Run Ollama as a companion service:

podman-compose --profile with-ollama up -d ollama

Note: If podman-compose errors with --userns and --pod cannot be set together, re-run with podman-compose --in-pod=false up. This is a known interaction between userns_mode: keep-id and podman-compose's default pod-per-project behavior on some versions.

Graphical Interface (folge_gui)

folge_gui is an accessible, browser-based front end for the pipeline, built with NiceGUI. It lives in src/folge_gui and is a parallel, additive companion to folge-cli — every command it runs is launched as a real subprocess (the same folge-cli you'd run from a terminal), and src/folge_cli is never modified or imported for execution. It targets WCAG 2.2 AA: real heading levels, skip links, labeled form controls, ARIA live status announcements, visible focus indicators, and non-dismissible dialogs for interactive prompts.

Install and run

git clone https://github.com/mrhunsaker/Folge_Accessibility.git
cd Folge_Accessibility
uv sync --all-packages   # one-time: pulls in NiceGUI as a workspace member
uv run folge-gui         # opens http://localhost:8765

folge-gui is registered as a uv workspace member, so it shares one uv.lock and one .venv with folge-cli. The folge_gui console script is kept as an alias — uv run folge_gui works identically. folge-cli itself is completely unaffected.

What it does

Page	Purpose
Setup	Check `uv`, Pandoc, `pdfinfo`, and `pymupdf`; review every provider's resolved settings; edit `.env` and `config.yaml` from the browser
Steps	Run any `folge-cli` sub-command with its own form, live status, streamed output, and a post-run quality-gate dialog
Full Pipeline	Run `folge-cli pipeline` end to end behind an 8-stage tracker, swapping the CLI's two terminal prompts for accessible dialogs

See docs/gui.md and src/folge_gui/README.md for the full details, including the contrast-ratio table for the Catppuccin Latte theme.

Folge_Accessibility/
├── pyproject.toml                  # Project metadata, dependencies, entry point
├── config.yaml                     # Provider settings, targets, validation
├── .env                            # Environment variables (provider, API keys)
├── run_pipeline.py                 # Backward-compatible entry point
│
├── src/folge_cli/                  # Installable Python package
│   ├── __init__.py                 # Package metadata
│   ├── __main__.py                 # python -m folge_cli
│   ├── _version.py                 # Dynamic version (CalVer)
│   ├── cli.py                      # folge-cli entry point (10 subcommands)
│   ├── config.py                   # Centralized configuration loading
│   ├── pipeline.py                 # Full pipeline orchestrator
│   ├── batch_process.py            # Vision API image processing
│   ├── merge.py                    # Deterministic merge (guide + vision)
│   ├── render.py                   # Markdown rendering via Jinja2
│   ├── publish.py                  # Standalone publisher with PDF/UA
│   ├── metadata.py                 # Accessible document metadata generator
│   ├── validate_schema.py          # JSON schema validation
│   ├── validate_content.py         # Content quality validation
│   ├── validate_pdf.py             # PDF/UA compliance validation
│   ├── generate_manual_attention.py # Manual attention markdown generation
│   └── progress.py                 # Step counters and progress display
│
├── src/folge_gui/                  # Accessible NiceGUI front end (workspace member)
│   ├── pyproject.toml              # folge-gui package + folge-gui/folge_gui scripts
│   ├── app.py                      # Route registration + ui.run()
│   ├── theme.py                    # Catppuccin Latte tokens, WCAG-safe "ink" variants
│   ├── a11y.py                     # headings, live regions, skip-link landmarks
│   ├── process_runner.py           # async subprocess runner, prompt detection
│   ├── prereqs.py                  # structured prerequisite / provider checks
│   ├── config_io.py                # .env / config.yaml read + write
│   ├── steps.py                    # form <-> argv mapping for each folge-cli command
│   ├── components.py               # step cards, console, status tracker, dialogs
│   └── pages/
│       ├── home.py                 # Landing page
│       ├── setup.py                # Prerequisites, provider settings, config editors
│       ├── steps_page.py           # Individual sub-command runner
│       └── pipeline_page.py        # Full pipeline with 8-stage tracker
│
├── pyinstaller/                    # PyInstaller build configuration
│   └── folge-cli.spec              # Spec file for building executables
│
├── scripts/                        # Backward-compatible thin shims
├── templates/                      # Jinja2 templates, Lua filters, CSS
│   ├── markdown.md                 # Jinja2 Markdown rendering template
│   ├── prompt.txt                  # Vision AI prompt template
│   ├── folge.css                   # @font-face declarations and base styles
│   ├── letter-portrait.css         # PDF @page layout for Letter portrait
│   └── letter-landscape.css        # PDF @page layout for Letter landscape
├── fonts/                          # Bundled accessible fonts
│   ├── Atkinson_Hyperlegible_Next/ # Variable weight text font
│   └── AtkinsonHyperlegibleMono/   # Static OTF monospace font
├── schemas/                        # JSON schemas for validation
├── docs/                           # MkDocs documentation source
├── images/                         # Container workflow only (docker-compose mounts)
├── output/                         # Container workflow only (docker-compose mounts)
│
├── .github/workflows/
│   ├── deploy-docs.yml             # Deploy MkDocs to GitHub Pages
│   └── release.yml                 # Build + release on v* tags
│
├── LICENSE                         # Apache 2.0
├── CHANGES.md                      # Changelog
├── CONTRIBUTING.md                 # How to contribute
├── STYLE.md                        # Code style guide
├── SECURITY.md                     # Security policy
├── GOVERNANCE_ENFORCEMENT.md       # Governance enforcement
└── CODE_OF_CONDUCT.md              # Community standards

Configuration Reference

File	Purpose
`pyproject.toml`	Project metadata, dependencies, `folge-cli` entry point, `[tool.uv.workspace]` members (`src/folge_gui`)
`config.yaml`	Provider settings, output targets (informational; registry in `folge_cli.formats`), validation thresholds, `project.author` and `project.keywords` metadata defaults
`.env`	Environment variables: provider selection, API keys, paths
`templates/prompt.txt`	Vision AI prompt template
`templates/markdown.md`	Jinja2 Markdown rendering template
`templates/folge.css`	Font declarations and base styles (Atkinson Hyperlegible)
`templates/letter-portrait.css`	PDF page layout for Letter portrait
`templates/letter-landscape.css`	PDF page layout for Letter landscape

Metadata Defaults

Setting	Default	Description
`project.author`	`Michael Ryan Hunsaker, M.Ed., Ph.D.`	Author fallback used by `folge-cli metadata`
`project.keywords`	`accessibility, documentation, ...`	Keyword fallbacks used by `folge-cli metadata`

Validation Settings

Setting	Default	Description
`MIN_CONFIDENCE`	`0.7`	Minimum confidence threshold for vision results
`require_alt_text`	`true`	Require alt text for all images
`require_long_description`	`true`	Require long descriptions for images
`max_alt_text_length`	`150`	Maximum characters for alt text

Contributing

We welcome contributions! Please see:

CONTRIBUTING.md — How to get started
STYLE.md — Code conventions and patterns
SECURITY.md — Vulnerability reporting process
GOVERNANCE_ENFORCEMENT.md — Enforcement policy
CODE_OF_CONDUCT.md — Community standards

License

This project is licensed under the Apache License, Version 2.0 — see the LICENSE file for details.

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Navigation Menu

Uh oh!

Folge Vision Publishing Pipeline

Overview

Input: `guide.json`

Input: `guide.enriched.json`

Pipeline Stages

Output Formats

Key Features

Pre-built Binaries

Quick Start

Prerequisites

Providers

Provider Table

Configuration

CLI Reference

Available Targets

Running Stages Individually

Accessible Document Metadata

Building from Source

Install

Build Executable

Container Usage

Docker

Podman

Graphical Interface (folge_gui)

Install and run

What it does

Configuration Reference

Metadata Defaults

Validation Settings

Contributing

License

Uh oh!

FilesExpand file tree

README.md

Latest commit

History

README.md

File metadata and controls

Folge Vision Publishing Pipeline

Overview

Input: guide.json

Input: guide.enriched.json

Pipeline Stages

Output Formats

Key Features

Pre-built Binaries

Quick Start

Prerequisites

Providers

Provider Table

Configuration

CLI Reference

Available Targets

Running Stages Individually

Accessible Document Metadata

Building from Source

Install

Build Executable

Container Usage

Docker

Podman

Graphical Interface (folge_gui)

Install and run

What it does

Configuration Reference

Metadata Defaults

Validation Settings

Contributing

License

Input: `guide.json`

Input: `guide.enriched.json`