Fullbleed PDF Engine

Create PDFs from HTML and CSS in Python, Rust, Node.js, C#, and browser JavaScript. MIT licensed.

Build invoices, reports, statements, and variable-data print jobs with a deterministic Rust engine. Use the native Rust API, the Python package, or the optional JavaScript and .NET integrations. The Python wheel includes the engine, fonts, and CLI with no required third-party Python packages or system PDF stack. The JavaScript package includes the WebAssembly engine and fonts for Node.js and browser workers; the .NET package includes native libraries for in-process rendering.

Try in your browser · Documentation · Examples · PyPI · crates.io · npm · Release notes · MIT license

Actual Fullbleed output. Open either preview for its PDF, or explore the designed examples with HTML/CSS, data, and font setup.

Choose your language

The Node.js quickstart and the JavaScript, React, Vue, and Next.js starters use package 0.4.1 with engine 2.5.22. They generate ordinary PDFs and previews. Import fullbleed in Node or fullbleed/browser in a web app. Node applications can use a bounded render queue and optional process isolation. FullBleed.DotNet 0.1.8 uses engine 2.5.22. Its managed library targets .NET 8 and works with .NET 8, 9, and 10 applications, with native rendering, compiled bindings, and previews. Its separate CLI adapter requires the Python CLI. The Python package provides the complete CLI, including compiled variable-data jobs and print profiles; the Rust guide covers the native API.

Your first PDF

Start in the HTML/CSS playground to edit a designed invoice or report and download a real PDF. It runs locally in your browser with no account or document upload. You can also use the Python notebook in Colab, or start with a local script:

Python

Python 3.10–3.14 on Windows, macOS, or Linux:

python -m pip install fullbleed

Save this as hello.py, then run python hello.py:

from pathlib import Path
import fullbleed

html = "<h1>Invoice INV-1042</h1><p>Consulting: USD 1,200.00</p>"
css = "@page { size: A4; margin: 20mm; } h1 { color: #175c52; }"
pdf = fullbleed.PdfEngine().render_pdf(html, css)
Path("invoice.pdf").write_bytes(pdf)

Open invoice.pdf. The quickstart adds fonts, previews, and a complete project. The runnable source is in examples/quickstart/hello.py.

Node.js

Use Node.js 22 or newer. Run npm install fullbleed, save this as invoice.mjs, then run node invoice.mjs:

import { writeFile } from 'node:fs/promises';
import { renderPdf } from 'fullbleed';

const result = await renderPdf({
  html: '<h1>Invoice INV-1042</h1><p>Consulting: USD 1,200.00</p>',
  css: '@page { size: A4; margin: 20mm } h1 { color: #175c52 }',
  previewDpi: 96,
});
await writeFile('invoice.pdf', result.pdf);
await writeFile('invoice.png', result.previews[0]);

Open invoice.pdf or its PNG preview. The Node.js quickstart includes a downloadable designed invoice project. The separate Node package supports ESM, CommonJS, and TypeScript. For an application route, use the Next.js starter.

Browser JavaScript

Try the editable browser app to generate an invoice or report, change its HTML/CSS, and download the PDF. Rendering runs locally in a Web Worker; the starter does not upload document inputs to a rendering service.

To build on it, download the complete project. Extract the ZIP, open a terminal in fullbleed-browser-starter, and use Node.js 22.12 or newer for the development tools:

npm ci
npm run dev

Open the localhost URL printed by Vite. The browser guide explains adding fullbleed/browser to an existing app, serving its runtime assets over HTTPS or localhost, cancellation, custom fonts, and static deployment. This renders supplied static HTML/CSS; it does not capture a live browser page.

Rust

Start with the native Rust quickstart. The fullbleed crate exposes FullBleed::builder() and render_to_buffer(...), with no Python requirement. The guide includes runnable invoice and report examples with explicit fonts and PNG previews, plus the latest hosted Rust API reference.

C# and .NET

Install FullBleed.DotNet in a .NET 8, 9, or 10 project and follow the C# quickstart. The guide includes a runnable invoice with explicit fonts and HTML/CSS source. Native rendering runs inside the .NET process on Windows x64, Linux x64, Intel macOS, and Apple Silicon macOS, without a Python or browser runtime.

Choose your starting point

See generated PDFs and their source →

Why use Fullbleed?

• Simple deployment: prebuilt wheels for 15 platform targets, with no browser or system-font installation needed for the bundled workflow.
• Repeatable documents: explicit assets and metadata, deterministic rendering, and SHA-256 reproducibility checks.
• Document-focused layout: pagination, headers and footers, tables, template overlays, and per-page data.
• Volume workflows: compile a template once, then bind fixed or reflowing records. Measured workloads and limits.
• Inspectable output: PDF inspection, PNG previews, structured diagnostics, and profile verification tooling.

Native previews of unembedded Standard 14 fonts use compact bundled outline substitutes, so basic Helvetica, Times, Courier, Symbol, and Dingbats text renders without system fonts. Registered and embedded font programs take precedence. The font-resolution trace identifies the bundled_substitute outcome; runtime discovery reports bundled_standard_font_previews. See the font sources, licenses, and coverage limits.

Fullbleed uses static HTML/CSS as a document layout language. Check the CSS coverage and remaining gaps for your templates. Use a browser when JavaScript or live website behavior is part of the required result.

Profile selection alone does not establish standards conformance. The release evidence records the checks performed and their limits.

Get help and contribute

Ask usage questions and share documents in Discussions. For a reproducible bug or feature proposal, open an issue. The contributor guide covers local setup, tests, and where to make changes.

If Fullbleed is useful to you, support its maintenance through sponsorship, documentation, or a useful example. Sponsorship is optional; Fullbleed remains MIT licensed and free to use.

Agent discovery and integrations

An agent does not need release-specific Fullbleed knowledge. The installed binary generates one canonical contract containing its actual version, selection boundary, capabilities, parser-derived CLI surface, result schemas, PDF profiles, examples, limitations, Agent Skill metadata, MCP tools, and acceptance tasks:

fullbleed agent-manifest --json
# Equivalent canonical form:
fullbleed agent-contract --format json

The committed fullbleed-agent-contract.json, cli_schema.md, and llms.txt are generated from the built wheel. CI rejects version, command, schema, profile, or capability drift. Runtime introspection remains authoritative.

The first-party, versionless skills/fullbleed/SKILL.md teaches tool selection and the render/preview/diagnose/verify loop without duplicating the manual. It is bundled in the wheel and can be exported into any absent or empty agent-skill directory:

fullbleed agent skill-path --json
fullbleed agent export-skill .agents/skills/fullbleed --json

You can also install the Skill directly from a pinned GitHub release with gh skill. The guide shows how to check its source and version; the Python package supplies the rendering engine.

For tool-calling agents, install the separately versioned adapter so the core package remains dependency-free:

python -m pip install fullbleed-mcp
fullbleed-mcp --root .

The stdio server confines document paths to the selected workspace and delegates every engine operation to the installed Fullbleed runtime. It exposes semantic discovery, project creation, render/preview, inspect, verify, assets, and fixed/reflow compiled VDP tools. The dependency-free core also provides fullbleed mcp --root ..

Use Fullbleed when the requested artifact is a deterministic print document built from structured content. Use a browser when the requested artifact is a screenshot or interactive state of an arbitrary live website. Use a general PDF editor when existing page content itself must be rewritten.

Cold-agent testing is available through fullbleed agent-acceptance; compact copyable workflows live in examples/agent_workflows, and the approach-neutral benchmark scaffold lives in agentdocbench.

This README is the canonical usage guide for:

• fullbleed CLI (human workflows + machine/agent automation)
• fullbleed Python bindings (PdfEngine, AssetBundle, batch APIs)

Additional focused references are in docs/:

• docs/install-non-technical.md (step-by-step setup for non-technical users)
• docs/css-coverage.md (validated CSS coverage, parity status, and active gaps)
• docs/README.md
• docs/engine.md
• docs/performance-architecture.md
• docs/performance-pass-2026-08-04.md
• docs/python-api.md
• docs/ui-accessibility.md
• docs/cli.md
• docs/pdf-templates.md

What You Get

• No headless browser requirement for PDF generation.
• Deterministic render pipeline with optional SHA256 output hashing.
• Reproducibility workflow via --repro-record and --repro-check.
• PDF 1.7 as the production-stable default target.
• Rust-native PDF template composition for VDP/transactional overlays.
• Native Rust image emission for overlay and finalized compose outputs (--emit-image) without external PDF raster runtime dependencies.
• Feature-driven page-to-template binding with per-page deterministic compose plans.
• Structured JSON result schemas for CI and AI agents.
• Offline-first asset model with explicit remote opt-in.
• Remote project template registry workflows (new list, new search, new remote).
• Python-first extension surface for hackability and custom workflows.
• Python render calls release the GIL while Rust rendering executes.
• Ordered standard-library worker pools for batch rendering and selected internal workloads.
• A deliberately small, license-audited Rust dependency graph and no required third-party Python runtime or build packages.

Concurrency Model

• Python binding render methods release the GIL during Rust execution through the Stable ABI bridge.
• Parallel batch APIs use Fullbleed's ordered standard-library worker pool (render_pdf_batch_parallel(...) and parallel-to-file variants).
• The same bounded worker implementation serves selected internal hotspots such as table layout and JIT paint paths.
• Do not assume every single-document render path will fully saturate all cores end-to-end.

Performance in 2.2

Fullbleed 2.2 retains exact TrueType glyph subsetting and adds compiled fixed-geometry variable-data binding. Across the five independent benchmark fixtures, PDFs are 95.5-97.3% smaller and ordinary warm rendering is 2.90x faster by geometric mean than released 2.0.0.

The new compile-once API freezes an immutable Q32.32 display document and can link it repeatedly without rerunning HTML parsing, selector matching, layout, pagination, or command planning. For identical untagged copies in one ordered PDF, CompiledDocument.render_pdf_batch(...) shares each source page's content stream across the copied page dictionaries. The measured 20-copy lane is 200.7x faster per page by geometric mean than the released 2.0.0 warm renderer; a 1,000-page stress run sustained 304,479-666,622 pages/s.

The 2.2 compiler includes a genuine fixed-geometry variable-data lane. A compiled invoice with six bound fields rendered 100,000 distinct records in a five-run median of 283.807 ms to memory (352,352 pages/s) and 295.999 ms directly to a flushed 88.1 MB PDF (337,839 pages/s). Including the one-time 4.168 ms compile gives 333,147 direct-file pages/s. The harness verified all 100,000 unique invoice IDs in page order, exact page count, resolved markers, deterministic bytes, and equal buffer/file SHA-256. Reproduce it with:

python tools/benchmark_fullbleed_vdp.py --records 100000 --repeats 5

Fixed-geometry slots may remain inside immutable page-space transforms and clips. The compiler captures that active coordinate-state program once and replays it in each compact record overlay; it does not rerun layout or clone the complete page display list.

Base-14/WinAnsi slot paint keeps the measured byte-patch linker. Registered embedded fonts use a command-level overlay program so each bound value is shaped and included in the font subset while the compiled layout, pagination, and static page stream remain shared. The throughput figures above describe the Base-14 fast lane; they are not claims for registered-font shaping.

The fixed-copy result remains scoped to identical content; the variable-data result is scoped to paint-only text whose geometry does not reflow. Neither is a claim that arbitrary new HTML or size-changing records render 200x faster. See docs/performance-pass-2026-08-04.md for measurements and docs/performance-architecture.md for the packed vector IR, typed-binding, virtual-linker, and shader roadmap.

Fullbleed 2.2.5 hardens the third compiler lane introduced in 2.2.4 for size-changing records: CompiledDocument.render_pdf_reflow_bindings(...). It parses and recovers the template DOM once and compiles encountered structural flow variants into guarded fixed-point display programs. Matching records bind directly into those programs; workers execute pre-shaped text/TJ paint slots, cached static PDF page segments, page-local Deflate, and one ordered linker. Explicit trusted structural slots use data-fb-bind-html="slot". A value that no compiled variant can safely place runs ordinary fixed-point layout once to add another variant.

On the independent 1,000-distinct-record case study, the 29-sample hot direct-file median was 216.160 ms: 4,626 records/s and 8,096 pages/s. The 168.995 ms best sample reached 5,917 records/s and 10,355 pages/s. It produced the exact 1,750-page 500/300/150/50 reflow distribution and verified all 24,900 markers. The throughput-tuned 6,870,320-byte PDF was deterministic across all samples. With the compact compression setting, compiled output was also byte-for-byte identical to ordinary rendering: 5,298,961 bytes and SHA-256 bb3c441313a08fb00d3bd15f23a567981f3bbbd550908e3a9fe7a07ca5d7f138. This is a hot compiled-variant result, not a claim for previously unseen structure. The first variant-discovery render measured 0.509 s (3,437 pages/s), and the fully cold job including setup, bindings, compilation, discovery, and write measured 0.588 s (2,975 pages/s). The independently observed hot-median speedup over the ordinary path was 43.8x.

Install

New to Python or setting up on a fresh machine? Start with docs/install-non-technical.md.

python -m pip install fullbleed

From a local wheel:

python -m pip install C:\path\to\fullbleed-2.5.23-cp310-abi3-win_amd64.whl

From a source checkout with Rust installed, no Python build package is needed:

python -m pip install --no-build-isolation --no-deps --editable .

To create deterministic release artifacts directly:

python build_backend/fullbleed_build_backend.py wheel --out dist
python build_backend/fullbleed_build_backend.py sdist --out dist

Platform artifact policy:

• Linux wheels cover manylinux2014 on x86-64, x86, ARM64, ARMv7, s390x, and ppc64le, plus musllinux_1_2 on x86-64, x86, ARM64, and ARMv7.
• Windows wheels cover x86-64, x86, and ARM64; macOS wheels cover Intel and Apple silicon.
• The CPython stable ABI allows each platform wheel to support Python 3.10 through 3.14.
• CI installs and exercises every built wheel on its target architecture, using native runners or QEMU as appropriate. The x86-64 manylinux wheel is additionally tested on every supported Python version before publication.

Separate Windows and Linux jobs check Python 3.15.0rc3 with the standard GIL, including the Python suite and byte-for-byte PDF/preview comparisons against Python 3.10. See the preview coverage and its limits.

Verify command surface:

fullbleed --help
fullbleed capabilities --json
fullbleed doctor --json

60-Second Quick Start (Project Happy Path)

Initialize project scaffold:

fullbleed init .

fullbleed init now vendors Bootstrap (5.0.0) into vendor/css/bootstrap.min.css, vendors Bootstrap Icons (1.11.3) into vendor/icons/bootstrap-icons.svg, vendors inter into vendor/fonts/Inter-Variable.ttf, writes license notices (vendor/css/LICENSE.bootstrap.txt, vendor/icons/LICENSE.bootstrap-icons.txt, vendor/fonts/LICENSE.inter.txt), and seeds assets.lock.json with pinned hashes. The scaffolded report.py also runs a component mount smoke validation before main render and writes output/component_mount_validation.json (fails fast on missing glyphs, actual pagination overflow, or CSS miss signals). On current engines, render-time pagination is authoritative for overflow; conservative JIT placement bounds remain the compatibility fallback for older engines. Scaffolded components now include components/primitives.py with reusable layout/content helpers (Stack, Row, Text, table/list helpers, key/value rows, etc.). Each scaffolded project also includes SCAFFOLDING.md, which should be your first read before restructuring components. It also includes AGENTS.md, which preserves the installed-runtime-first document workflow across future coding-agent sessions without making other document tools categorically forbidden.

Install additional project assets (defaults to ./vendor/... in project context):

fullbleed assets install inter --json

Bootstrap baseline note:

• We target Bootstrap (5.0.0) as the default styling baseline for project workflows.
• Re-run fullbleed assets install bootstrap --json only if you want to explicitly refresh/bootstrap-manage outside init.

Render using the scaffolded component pipeline:

python report.py

Expected artifacts from scaffolded report.py:

• output/report.pdf
• output/report_page1.png (or equivalent page preview from engine image APIs)
• output/component_mount_validation.json
• output/css_layers.json

Canonical static PDF reference:

python examples/canonical_reference/report.py

examples/canonical_reference/ is the exhaustive scaffold-shaped reference for component composition, layered CSS, bundled fonts/SVG, inline SVG, raster data URIs, linked and standalone HTML artifacts, PDF output, PNG previews, and validation reports.

Project Bootstrap Templates (fullbleed new)

Use local starters:

fullbleed new local invoice ./my-invoice
fullbleed new local statement ./my-statement
fullbleed new local accessible ./my-accessible-doc
fullbleed new local reference ./my-reference-doc

Discover remote starters from registry:

fullbleed new list --json
fullbleed new search i9 --tag vdp --json
fullbleed new remote i9-stamped-vdp ./i9-job --json

fullbleed new local accessible is the verbose accessibility-first starter and demonstrates the fullbleed.accessibility runtime surface (engine verifier, PMR, PDF/UA-targeted seed checks, and non-visual trace artifacts). fullbleed new local reference vendors the canonical static PDF reference shape as a scaffolded project with component layers, assets, PDF/PNG outputs, page data, and validation reports.

Optional registry override (for private/canary registries):

fullbleed new list --registry https://example.com/manifest.json --json

or:

set FULLBLEED_TEMPLATE_REGISTRY=https://example.com/manifest.json
fullbleed new search statement --json

Scaffold-First Workflow (Recommended)

fullbleed init is designed for component-first authoring rather than a single large HTML template.

Typical scaffold layout:

.
|-- SCAFFOLDING.md
|-- COMPLIANCE.md
|-- report.py
|-- components/
|   |-- fb_ui.py
|   |-- primitives.py
|   |-- header.py
|   |-- body.py
|   |-- footer.py
|   `-- styles/
|       |-- primitives.css
|       |-- header.css
|       |-- body.css
|       `-- footer.css
|-- styles/
|   |-- tokens.css
|   `-- report.css
|-- vendor/
|   |-- css/
|   |-- fonts/
|   `-- icons/
`-- output/

Best-practice authoring model:

1. Read SCAFFOLDING.md first for project conventions.
2. Keep composition and data loading in report.py.
3. Keep reusable component building blocks in components/primitives.py.
4. Keep section markup in components/header.py, components/body.py, components/footer.py.
5. Keep component-local styles in components/styles/*.css.
6. Keep page tokens/composition styles in styles/tokens.css and styles/report.css.

Recommended CSS layer order:

1. styles/tokens.css
2. components/styles/primitives.css
3. components/styles/header.css
4. components/styles/body.css
5. components/styles/footer.css
6. styles/report.css

Recommended iteration loop:

1. Edit data loading + component props in report.py.
2. Edit component markup in components/*.py.
3. Edit styles in components/styles/*.css and styles/*.css.
4. Run python report.py.
5. Review output/report_page1.png, output/component_mount_validation.json, and output/css_layers.json.

Optional scaffold diagnostics:

• FULLBLEED_DEBUG=1 to emit JIT traces.
• FULLBLEED_PERF=1 to emit perf traces.
• FULLBLEED_EMIT_PAGE_DATA=1 to persist page data JSON.
• FULLBLEED_IMAGE_DPI=144 (or higher) for preview resolution.
• FULLBLEED_VALIDATE_STRICT=1 for stricter validation gates in CI.

One-off Quick Render (No Project Scaffold)

Render inline HTML/CSS with reproducibility artifacts:

fullbleed --json render \
  --html-str "<html><body><h1>Hello</h1></body></html>" \
  --css-str "body{font-family:sans-serif}" \
  --emit-manifest build/render.manifest.json \
  --emit-jit build/render.jit.jsonl \
  --emit-perf build/render.perf.jsonl \
  --deterministic-hash build/render.sha256 \
  --repro-record build/render.repro.json \
  --out output/hello.pdf

--deterministic-hash writes the output PDF SHA-256 by default; when --emit-image is enabled, it writes an artifact-set digest (fullbleed.artifact_digest.v1) over PDF SHA-256 plus ordered page-image SHA-256 hashes. JSON outputs expose outputs.deterministic_hash_mode (pdf_only or artifact_set_v1), with outputs.artifact_sha256 and outputs.image_sha256 when images are emitted.

Re-run and enforce reproducibility against a stored record:

fullbleed --json render \
  --html templates/report.html \
  --css templates/report.css \
  --repro-check build/render.repro.json \
  --out output/report.rerun.pdf

Generate PNG page artifacts from an existing validation render:

fullbleed --json verify \
  --html templates/report.html \
  --css templates/report.css \
  --emit-pdf output/report.verify.pdf \
  --emit-image output/report_verify_pages \
  --image-dpi 200

Compile-only plan (no render):

fullbleed --json plan \
  --html templates/report.html \
  --css templates/report.css

Template compose planning (no finalize write):

fullbleed --json plan \
  --html templates/overlay.html \
  --css templates/overlay.css \
  --template-binding config/template_binding.json \
  --templates config/template_catalog.json \
  --emit-compose-plan output/compose_plan.json

PDF Template Composition (VDP / Transactional)

When overlaying variable data onto a source PDF, use the built-in Rust template compose path.

Minimal CLI auto-compose flow:

fullbleed --json render \
  --html templates/overlay.html \
  --css templates/overlay.css \
  --asset templates/source.pdf --asset-kind pdf --asset-name source-template \
  --template-binding config/template_binding.json \
  --templates config/template_catalog.json \
  --out output/composed.pdf

Compose image semantics:

• In template auto-compose mode, --emit-image PNGs are rasterized from finalized composed pages and report outputs.image_mode=composed_pdf.
• In non-compose render/verify runs, --emit-image reports outputs.image_mode=overlay_document.

Minimal template_binding example:

{
  "default_template_id": "source-template",
  "feature_prefix": "fb.feature.",
  "by_feature": {
    "front": "source-template",
    "back_blank": "source-template"
  }
}

Python API compose flow:

import fullbleed

engine = fullbleed.PdfEngine(template_binding=binding_spec)
overlay_bytes, _page_data, _bindings = engine.render_pdf_with_page_data_and_template_bindings(html, css)
open("output/overlay.pdf", "wb").write(overlay_bytes)
plan_result = engine.plan_template_compose(
    html,
    css,
    [("source-template", "templates/source.pdf")],
    0.0,
    0.0,
)
plan = [
    (
        row["template_id"],
        row["template_page"],
        row["overlay_page"],
        row["dx"],
        row["dy"],
    )
    for row in plan_result["plan"]
]
fullbleed.finalize_compose_pdf(
    [("source-template", "templates/source.pdf")],
    plan,
    "output/overlay.pdf",
    "output/composed.pdf",
    annotation_mode="link_only",  # optional: link_only | none | carry_widgets
)

See docs/pdf-templates.md and examples/template-flagging-smoke/ for full production examples.

CLI Command Map

Schema discovery for any command/subcommand:

fullbleed --schema render
fullbleed --schema assets verify
fullbleed --schema inspect pdf
fullbleed --schema inspect templates

CLI Flags That Matter Most

Global machine flags:

• --json: structured result payload to stdout
• --json-only: implies --json and --no-prompts
• --schema: emit schema definition and exit
• --no-prompts: disable interactive prompts
• --config: load defaults from a config file
• --log-level error|warn|info|debug: control CLI log verbosity
• --no-color: disable ANSI color output
• --version: print CLI version and exit

Render/verify/plan key flags:

• Inputs: --html, --html-str, --css, --css-str --html accepts .svg files for direct SVG-document rendering; --html-str accepts inline SVG markup.
• Page setup: --page-size, --page-width, --page-height, --margin, --page-margins
• Engine toggles: --reuse-xobjects, --svg-form-xobjects, --svg-raster-fallback, --shape-text, --unicode-support, --unicode-metrics
• PDF/compliance: --pdf-version, --pdf-profile, --color-space, --document-lang, --document-title Stable default is --pdf-version 1.7 for shipping workflows. Profile targets: none, pdfa1a, pdfa1b, pdfa2a, pdfa2b, pdfa2u, pdfa3a, pdfa3b, pdfa3u, pdfa4, pdfa4e, pdfa4f, pdfx4, pdfua1, pdfua2, pdfvt1, wtpdf1r, wtpdf1a, tagged. Aliases: a, ua, vt, wt1r, wt1a, pdf/a, pdf/ua, pdf/vt. Output intent metadata (--output-intent-identifier|--output-intent-info|--output-intent-components) requires --output-intent-icc. pdfx4 and pdfvt1 force PDF 1.6 and require a title and explicit write date (--timestamp current, a fixed UTC date, or --timestamp-source SOURCE_DATE_EPOCH). PDF/VT composition documents record hierarchies, --pdf-vt-job, private DPM, conservative reuse hints, and the independent release-validation boundary. Run python tools/validate_pdf_profiles.py --download-verapdf --install-pdf-oxide --strict-external to regenerate profile specimens, capture inspect/JIT evidence, replay deterministic hashes, validate PDF/A and PDF/UA with veraPDF, and validate PDF/X-4 with pdf_oxide. WTPDF profiles are validated with veraPDF wt1r/wt1a and include PDF Declaration evidence. pdfa4f emits and checks an associated EmbeddedFiles name tree. pdfvt1 also emits and checks a parsed DPart graph (DPartRoot, DPartRootNode, NodeNameList [/Job /Record /Document], leaf page range, and page /DPart references), including a supplemental multipage specimen for /Start and /End, reported as granular booleans plus pdfvt_dpart_graph_valid; use a dedicated PDF/VT preflight tool for third-party PDF/VT certification, or wire one into the same harness with --pdfvt-cmd "tool --input {pdf}" --pdfvt-version-cmd "tool --version" --require-dedicated-pdfvt.
• Watermarking: --watermark-text, --watermark-html, --watermark-image, --watermark-layer, --watermark-semantics, --watermark-opacity, --watermark-rotation
• Artifacts: --emit-jit, --emit-perf, --emit-glyph-report, --emit-page-data, --emit-compose-plan, --emit-image, --image-dpi, --deterministic-hash
• Assets: --asset, --asset-kind, --asset-name, --asset-trusted, --allow-remote-assets
• Profiles: --profile dev|preflight|prod
• Fail policy: --fail-on overflow|missing-glyphs|font-subst|budget
• Fallback policy: --allow-fallbacks (keeps fallback diagnostics, but does not fail missing-glyphs / font-subst gates)
• Reproducibility: --repro-record <path>, --repro-check <path>
• Budget thresholds: --budget-max-pages, --budget-max-bytes, --budget-max-ms
• Release gates: doctor --strict, compliance --strict --max-audit-age-days <n>

SVG Workflows

Fullbleed supports SVG in three practical CLI paths:

• Direct SVG document render via --html <file.svg>
• Inline SVG markup via --html-str "<svg ...>...</svg>"
• Referenced SVG assets via --asset <file.svg> (kind auto-infers to svg)

Standalone SVG file to PDF:

fullbleed --json render \
  --html artwork/badge.svg \
  --out output/badge.pdf

Inline SVG markup to PDF:

fullbleed --json render \
  --html-str "<svg xmlns='http://www.w3.org/2000/svg' width='200' height='80'><rect width='200' height='80' fill='#0d6efd'/><text x='16' y='48' fill='white'>Hello SVG</text></svg>" \
  --out output/inline-svg.pdf

HTML template with explicit SVG asset registration:

fullbleed --json render \
  --html templates/report.html \
  --css templates/report.css \
  --asset assets/logo.svg \
  --asset-kind svg \
  --out output/report.pdf

SVG render behavior flags:

• --svg-form-xobjects / --no-svg-form-xobjects
• --svg-raster-fallback / --no-svg-raster-fallback

Distributed Python wheels enable the svg_raster engine feature, so --svg-raster-fallback can rasterize unsupported SVG constructs such as SVG text, filters, masks, and foreignObject into deterministic image content. Custom source builds must include --features python,svg_raster to advertise and use that fallback path.

Machine discovery:

fullbleed capabilities --json

Inspect the svg object in fullbleed.capabilities.v1 for SVG support metadata. The engine object also reports compiled-document/reflow availability and supported per-call compression modes. It reports the compiled svg_raster build feature plus a feature matrix for native-vector SVG, raster-fallback-required SVG, and unsupported/known-loss SVG features.

Image Support Matrix

Launch-safe image claims:

• Supported direct raster inputs: PNG and JPEG.
• SVG is handled by the SVG pipeline described above, with native-vector output where supported and raster fallback for fallback-only features when svg_raster is enabled.
• Supported references include filesystem paths, registered bundle assets, file/data URIs, <img>, CSS background-image: url(...), list-style images, and watermark images.
• Unsupported or not launch-claimed as direct inputs: WebP, GIF, TIFF, AVIF, BMP, animated images, <picture>, srcset, sizes, density descriptors, and browser-style responsive image selection.
• For deterministic builds, prefer vendored local assets or registered AssetBundle inputs; remote assets must be explicitly allowed.

Per-Page Templates (page_1, page_2, page_n)

Fullbleed uses ordered page templates internally. In docs, this is easiest to think of as:

• page_1: first page template
• page_2: second page template
• page_n: repeating template for later pages

Configuration mapping:

• CLI --page-margins keys: 1, 2, ... and optional "n" (or "each" alias).
• Python PdfEngine(page_margins=...): same key model.
• Missing numeric pages fall back to the base margin.
• The last configured template repeats for remaining pages.

Minimal CLI example:

{
  "1": {"top": "12mm", "right": "12mm", "bottom": "12mm", "left": "12mm"},
  "2": {"top": "24mm", "right": "12mm", "bottom": "12mm", "left": "12mm"},
  "n": {"top": "30mm", "right": "12mm", "bottom": "12mm", "left": "12mm"}
}

fullbleed --json render \
  --html templates/report.html \
  --css templates/report.css \
  --page-margins page_margins.json \
  --header-each "Statement continued - Page {page} of {pages}" \
  --out output/report.pdf

Minimal Python example:

import fullbleed

engine = fullbleed.PdfEngine(
    page_width="8.5in",
    page_height="11in",
    margin="12mm",
    page_margins={
        1: {"top": "12mm", "right": "12mm", "bottom": "12mm", "left": "12mm"},  # page_1
        2: {"top": "24mm", "right": "12mm", "bottom": "12mm", "left": "12mm"},  # page_2
        "n": {"top": "30mm", "right": "12mm", "bottom": "12mm", "left": "12mm"} # page_n
    },
    header_first="Account Statement",
    header_each="Statement continued - Page {page} of {pages}",
    footer_last="Final page",
)

Note:

• CLI currently exposes --header-each / --footer-each (and --header-html-each / --footer-html-each).
• For first/last header/footer variants (header_first, header_last, footer_first, footer_last), use the Python API.

Asset Workflow (CLI)

List installed + available packages:

fullbleed assets list --available --json

Install builtin assets:

fullbleed assets install bootstrap
fullbleed assets install bootstrap-icons
# `@bootstrap` / `@bootstrap-icons` are also supported aliases

PowerShell note:

• Quote @ aliases (for example "@bootstrap") to avoid shell parsing surprises.

Install remote asset package:

fullbleed assets install inter

Install broad Unicode fallback package (larger font payload):

fullbleed assets install noto-sans

Install to a custom vendor directory:

fullbleed assets install bootstrap --vendor ./vendor

Install to global cache:

fullbleed assets install inter --global

Install common barcode fonts (license-safe defaults):

fullbleed assets install libre-barcode-128
fullbleed assets install libre-barcode-39
fullbleed assets install libre-barcode-ean13-text

Verify against lock file with strict failure:

fullbleed assets verify inter --lock --strict --json

Preview cache cleanup without deleting files:

fullbleed cache prune --max-age-days 30 --dry-run --json

Notes:

• Builtin packages accept both plain and @ references (bootstrap == @bootstrap, bootstrap-icons == @bootstrap-icons, noto-sans == @noto-sans).
• noto-sans is available as a builtin fallback package, but it is intentionally larger than inter; use it when your document requires broader glyph coverage.
• Project installs default to ./vendor/ when project markers are present (assets.lock.json, report.py, or fullbleed.toml in CWD).
• If no project markers are found, assets install defaults to global cache unless --vendor is explicitly set.
• Do not hardcode cache paths like %LOCALAPPDATA%/fullbleed/cache/...; use assets install --json and consume installed_to.
• Installed assets include license files in typed vendor directories (for example vendor/fonts/, vendor/css/).
• assets lock --add is currently aimed at builtin package additions.
• Barcode packages in the remote registry are currently OFL-1.1 families from Google Fonts (Libre Barcode).
• USPS IMB fonts are not currently auto-installable via assets install; use local vetted font files and track licensing separately.

Bootstrap Vendoring

Bootstrap builtin package details:

• Package: bootstrap (alias: @bootstrap)
• Bundled version: 5.0.0
• Asset kind: CSS (bootstrap.min.css)
• Default install location: vendor/css/bootstrap.min.css (project mode)
• License: MIT
• License source: https://raw.githubusercontent.com/twbs/bootstrap/v5.0.0/LICENSE

Bootstrap Icons builtin package details:

• Package: bootstrap-icons (alias: @bootstrap-icons)
• Bundled version: 1.11.3
• Asset kind: SVG sprite (bootstrap-icons.svg)
• Default install location: vendor/icons/bootstrap-icons.svg (project mode)
• License: MIT
• License source: https://raw.githubusercontent.com/twbs/icons/v1.11.3/LICENSE

Bootstrap notes:

• Bootstrap is vendored and installable through the asset pipeline.
• Bootstrap CSS is consumed as an explicit local asset (--asset @bootstrap or AssetBundle); external HTML <link rel="stylesheet"> is not an execution path.
• Bootstrap preflight examples remain useful smoke examples, but they are not the canonical source of CSS parity claims.

Validated CSS Coverage

For the full, maintained coverage statement, see docs/css-coverage.md.

Summary as of May 19, 2026:

• Tracked CSS modules: 22
• Module state: 22/22 in_progress
• Full CSS fixture lane: 86/86 fixtures passing with 603 assertion/paint checks
• Parity status check: green (tools/generate_css_parity_status.py --check --json)
• Canonical validation artifact: _css_working/css_parity_status.json
• Canonical validation artifact: _css_working/tmp/fixture_full_latest.json
• Canonical validation artifact: _css_working/css_broad_coverage_sprint_s14.md

The complete pinned independent Wheel265 gate closed on August 9, 2026 with 1,642 PASS, 0 FAIL, and 20 REFERENCE-DISPUTED across 1,662 fixtures. That is strict 100% parity for every adjudicable fixture in this named corpus; disputed references remain visible and are not relabeled as implementation passes. Broader web-platform coverage remains bounded by the module and known-gap contract in docs/css-coverage.md.

Flow/reflow hardening validated on August 5, 2026 adds fixed-height block, one-column grid, and column-flex continuation slicing; avoids content duplication; retains empty styled captions inside avoided figures; preserves display: contents grid paint phase; and keeps positioned containing blocks atomic. The independent review's targeted 18-case regression cluster passes 18/18 (13 pixel-exact, maximum above-floor difference 0.97030199%). This targeted diagnostic does not replace the full external-corpus report.

Current validated behavior includes first-class parser -> evaluator -> calculator -> layout/paint coverage across broad static-document CSS domains (values math, layout primitives, pagination/fragmentation baselines, transforms phase-1, gradient/effects subsets including color-first and interleaved-color filter: drop-shadow(...) with computed lengths, modern rgb() shadow colors with mm lengths, duplicate drop-shadow(...) color rejection, empty optional filter-function defaults, explicit/currentColor backdrop-filter: drop-shadow(...) raster paint, strict box-shadow length/color grammar with modern rgb() plus mm paint coverage, softened blurred inset edge paint, and directional inset offset edge paint, overflow-gated text-overflow: ellipsis, physical min/max box constraints, horizontal-tb RTL direction inline inset/margin/padding/border mapping, vertical writing-mode direction: rtl inline inset remapping, vertical-rl logical sizing/insets/margin/padding/border/min-max constraints, vertical-lr logical sizing/min-max/insets/margin/padding/border remapping, basic vertical text columns with vertical-rl leftward and vertical-lr rightward line progression, and deterministic diagnostics).

Known gap categories are explicitly tracked for the final parity push:

• remaining filter/backdrop-filter function breadth beyond current subset, including SVG url() filters, advanced drop-shadow() forms beyond the current flexible-color-order foreground and baseline backdrop coverage, and PDF-native foreground filters
• advanced clip-path grammar beyond the current basic-shape subset, including SVG clip sources, SVG-specific geometry boxes, and deeper edge grammar
• plus-lighter PDF-native parity and deeper nested compositing/isolation breadth
• deeper blurred inset shadow edge fidelity
• remaining multi-layer background image edge semantics beyond the current sized gradient-layer repeat-x/space/round and PNG url(...) explicit-size/auto-auto-intrinsic/negative-size-invalid/contain/cover/auto-dimension-size/single-value-size/alpha-stack/default-repeat-repeat/repeat-y/repeat/repeat-no-repeat/no-repeat-repeat/space-space/round-round/round-space/space-round/round-repeat/round-no-repeat/no-repeat-round/no-repeat-space/space-no-repeat/repeat-space/repeat-round/space-repeat/logical-repeat-aliases/multi-layer-repeat-axis/background-list-repetition/percentage-position/edge-offset-position/logical-position-aliases/shorthand-position-size-repeat/content-box-origin-clip/multi-layer-origin-clip/background-blend-normal/background-blend-multiply/screen-mode/overlay-mode/exclusion-mode/hard-light-mode/darken-mode/lighten-mode/color-dodge-mode/color-burn-mode/soft-light-mode/hue-mode/saturation-mode/color-mode/luminosity-mode/plus-lighter-mode/list-repetition/truncation/raster-blend/raster-layer-mapping/mixed-raster-gradient-mapping/mixed-gradient-raster-mapping subset, including broader mixed raster stack combinations
• table layout edge semantics hardening beyond the current fixed-layout width-hint, auto-width, caption-side placement, and invalid caption-side inheritance lanes
• full vertical text/layout flow beyond the current vertical static-layout and basic wrapped text-column baseline

run Command (Python Factory Interop)

run lets the CLI use a Python-created engine instance.

report.py:

import fullbleed

def create_engine():
    return fullbleed.PdfEngine(page_width="8.5in", page_height="11in", margin="0.5in")

CLI invocation:

fullbleed --json run report:create_engine \
  --html-str "<h1>From run</h1>" \
  --css templates/report.css \
  --out output/report.pdf

Entrypoint formats:

• module_name:factory_or_engine
• path/to/file.py:factory_or_engine

Rust chart compilation

For Rust-based visual/source frontends, the engine also provides document-context chart compilation: normal CSS font selection and size, project-font legend measurement/wrapping, semantic tables, and observable layout failures. Generated SVG is derived output, not authoring source.

Python API Quick Start

import fullbleed

engine = fullbleed.PdfEngine(
    page_width="8.5in",
    page_height="11in",
    margin="0.5in",
    pdf_version="1.7",
    pdf_profile="none",
    color_space="rgb",
)

html = "<html><body><h1>Invoice</h1><p>Hello.</p></body></html>"
css = "body { font-family: sans-serif; }"

bytes_written = engine.render_pdf_to_file(html, css, "output/invoice.pdf")
print(bytes_written)

Register local assets with AssetBundle:

import fullbleed

bundle = fullbleed.AssetBundle()
bundle.add_file("vendor/css/bootstrap.min.css", "css", name="bootstrap")
bundle.add_file("vendor/fonts/Inter-Variable.ttf", "font", name="inter")

engine = fullbleed.PdfEngine(page_width="8.5in", page_height="11in")
engine.register_bundle(bundle)
engine.render_pdf_to_file("<h1>Styled</h1>", "", "output/styled.pdf")

Python API Signatures (Runtime-Verified)

These signatures are verified from the installed package via inspect.signature(...).

PdfEngine constructor:

PdfEngine(
    page_width=None,
    page_height=None,
    margin=None,
    page_margins=None,
    font_dirs=None,
    font_files=None,
    reuse_xobjects=True,
    svg_form_xobjects=False,
    svg_raster_fallback=False,
    unicode_support=True,
    shape_text=True,
    unicode_metrics=True,
    pdf_version=None,
    pdf_profile=None,
    output_intent_icc=None,
    output_intent_identifier=None,
    output_intent_info=None,
    output_intent_components=None,
    color_space=None,
    document_lang=None,
    document_title=None,
    header_first=None,
    header_each=None,
    header_last=None,
    header_x=None,
    header_y_from_top=None,
    header_font_name=None,
    header_font_size=None,
    header_color=None,
    header_html_first=None,
    header_html_each=None,
    header_html_last=None,
    header_html_x=None,
    header_html_y_from_top=None,
    header_html_width=None,
    header_html_height=None,
    footer_first=None,
    footer_each=None,
    footer_last=None,
    footer_x=None,
    footer_y_from_bottom=None,
    footer_font_name=None,
    footer_font_size=None,
    footer_color=None,
    watermark=None,
    watermark_text=None,
    watermark_html=None,
    watermark_image=None,
    watermark_layer="overlay",
    watermark_semantics="artifact",
    watermark_opacity=0.15,
    watermark_rotation=0.0,
    watermark_font_name=None,
    watermark_font_size=None,
    watermark_color=None,
    paginated_context=None,
    jit_mode=None,
    debug=False,
    debug_out=None,
    perf=False,
    perf_out=None,
)

For regular/italic family selection and exact face mappings, see Registering fonts.

Module exports:

• PdfEngine
• CompiledDocument
• CompiledFlowCompression
• AssetBundle
• Asset
• AssetKind
• WatermarkSpec(kind, value, layer='overlay', semantics=None, opacity=0.15, rotation_deg=0.0, font_name=None, font_size=None, color=None)
• concat_css(parts)
• vendored_asset(source, kind, name=None, trusted=False, remote=False)
• inspect_pdf(path)
• inspect_template_catalog(templates)
• finalize_stamp_pdf(template, overlay, out, page_map=None, dx=0.0, dy=0.0)
• finalize_compose_pdf(templates, plan, overlay, out, annotation_mode='link_only')
• fetch_asset(url)

PdfEngine methods:

CompiledDocument methods:

The compiled lane runs parsing/layout once. render_pdf_batch virtualizes identical untagged page content. The fixed binding methods accept columnar dict[str, list[str]] data and emit distinct fixed-geometry text overlays per record while sharing static paint. They require exact slot keys, equal non-zero column lengths, page-local text, and an untagged profile. Base-14 text uses the direct byte-patch lane; registered fonts are shaped into a command-level overlay. Bound values do not trigger layout or reflow.

The reflow binding methods use the same exact-column contract but execute the compiled parsed-DOM program. {{slot}} values are literal text and may reshape, wrap, resize blocks, and change page counts. Empty elements can opt into trusted generated markup with data-fb-bind-html="slot"; those values replace the element's children. One DOM is reused per native worker and record output is linked in input order through a bounded queue. Structural slot values are parsed as trusted markup: sanitize untrusted input or use scalar {{slot}} bindings, which remain literal text. Choose CompiledFlowCompression.Throughput (default) or CompiledFlowCompression.Compact per render; modes can be mixed safely in one process. See docs/python-api.md for the trust boundary and current limitations.

When deterministic_hash is set, engine writes PDF SHA-256 to the provided file path.

AssetBundle methods:

• add_file(path, kind, name=None, trusted=False, remote=False)
• add(asset)
• assets_info()
• css()

Python Examples (Smoke-Checked)

Canonical scaffold reference:

python examples/canonical_reference/report.py

Text watermark + diagnostics:

import fullbleed

engine = fullbleed.PdfEngine(
    page_width="8.5in",
    page_height="11in",
    margin="0.5in",
    pdf_version="1.7",
    watermark_text="INTERNAL",
    watermark_layer="overlay",
    watermark_semantics="artifact",
    watermark_opacity=0.12,
    watermark_rotation=-32.0,
    debug=True,
    debug_out="build/invoice.jit.jsonl",
    perf=True,
    perf_out="build/invoice.perf.jsonl",
)

html = "<h1>Invoice</h1><p>Status: Ready</p>"
css = "h1{margin:0 0 8px 0} p{margin:0}"

written = engine.render_pdf_to_file(html, css, "output/invoice_watermarked.pdf")
print("bytes:", written)

Batch render + glyph/page-data checks:

import fullbleed

engine = fullbleed.PdfEngine(page_width="8.5in", page_height="11in", margin="0.5in")

jobs = [
    ("<h1>Batch A</h1><p>Alpha</p>", "h1{color:#0d6efd}"),
    ("<h1>Batch B</h1><p>Beta</p>", "h1{color:#198754}"),
]

written = engine.render_pdf_batch_with_css_to_file(jobs, "output/batch.pdf")
print("batch bytes:", written)

pdf_bytes, glyph_report = engine.render_pdf_with_glyph_report("<p>Hello</p>", "")
print("glyph entries:", len(glyph_report))

pdf_bytes, page_data = engine.render_pdf_with_page_data("<p>Hello</p>", "")
print("page data available:", page_data is not None)

Transactional Header/Footer + Totals

Minimal, self-contained Python example (no external template files) showing:

• Continued headers on page 2+.
• Per-page subtotal footer expansion via {sum:items.amount}.
• Final-page grand total footer expansion via {total:items.amount}.
• Structured page_data totals for automation and reconciliation checks.

from pathlib import Path
import fullbleed

rows = []
for i in range(1, 121):  # enough rows to force multiple pages
    amount = 10.00 + ((i * 7) % 23) + 0.25
    rows.append(
        f'<tr data-fb="items.amount={amount:.2f}">'
        f"<td>2026-01-{(i % 28) + 1:02d}</td>"
        f"<td>Txn {i:03d}</td>"
        f'<td class="num">${amount:.2f}</td>'
        "</tr>"
    )

html = f"""<!doctype html>
<html>
<body>
  <h1>Monthly Statement</h1>
  <table>
    <thead>
      <tr><th>Date</th><th>Description</th><th class="num">Amount</th></tr>
    </thead>
    <tbody>
      {''.join(rows)}
    </tbody>
  </table>
</body>
</html>
"""

css = """
body { font-family: sans-serif; font-size: 10pt; color: #111; }
h1 { margin: 0 0 8pt 0; }
table { width: 100%; border-collapse: collapse; }
th, td { padding: 4pt; border-bottom: 1pt solid #e1e1e1; }
thead th { background: #f3f6fa; text-transform: uppercase; font-size: 9pt; }
.num { text-align: right; }
"""

engine = fullbleed.PdfEngine(
    page_width="8.5in",
    page_height="11in",
    margin="12mm",
    page_margins={
        1: {"top": "12mm", "right": "12mm", "bottom": "12mm", "left": "12mm"},
        2: {"top": "28mm", "right": "12mm", "bottom": "12mm", "left": "12mm"},
        "n": {"top": "28mm", "right": "12mm", "bottom": "12mm", "left": "12mm"},
    },
    header_html_each=(
        '<div style="display:flex;justify-content:space-between;border-bottom:1pt solid #d9d9d9;">'
        '<div style="font-weight:bold;">Acme Ledger</div>'
        '<div style="font-size:9pt;color:#444;">Statement Continued - Page {page} of {pages}</div>'
        "</div>"
    ),
    header_html_x="12mm",
    header_html_y_from_top="6mm",
    header_html_width="186mm",
    header_html_height="10mm",
    paginated_context={"items.amount": "sum"},
    footer_each="Subtotal (Page {page}): ${sum:items.amount}",
    footer_last="Grand Total: ${total:items.amount}",
    footer_x="12mm",
    footer_y_from_bottom="8mm",
)

pdf_bytes, page_data = engine.render_pdf_with_page_data(html, css)
Path("output_transactional_minimal.pdf").write_bytes(pdf_bytes)

assert page_data["page_count"] >= 2
assert page_data["totals"]["items.amount"]["value"] == sum(
    p["items.amount"]["value"] for p in page_data["pages"]
)
print("Wrote output_transactional_minimal.pdf")
print("Page count:", page_data["page_count"])
print("Grand total:", page_data["totals"]["items.amount"]["formatted"])

API note:

• For transactional running totals (paginated_context) and HTML header/footer placement (header_html_*, footer_html_*), use the Python PdfEngine API path.
• The CLI currently exposes direct text header/footer flags (--header-each, --footer-each) for simpler cases.

CLI watermark parity example:

fullbleed --json render \
  --html-str "<h1>Watermark probe</h1><p>hello</p>" \
  --css-str "body{font-family:sans-serif}" \
  --watermark-text "INTERNAL" \
  --watermark-layer overlay \
  --watermark-opacity 0.12 \
  --watermark-rotation -32 \
  --out output/watermark_probe.pdf

Reference-Image Parity Workflow (Practical)

When targeting a design reference image (for example reference image exports), this loop has worked well:

1. Start from fullbleed init so CSS/font/icon baselines are vendored and pinned.
2. For scaffolded projects, run python report.py and set FULLBLEED_IMAGE_DPI as needed for sharper previews.
3. For direct CLI template rendering, register assets through the CLI (--asset ...) or AssetBundle.
4. Iterate with image artifacts enabled:

fullbleed --json render \
  --profile preflight \
  --html templates/invoice.html \
  --css templates/invoice.css \
  --asset vendor/css/bootstrap.min.css --asset-kind css --asset-name bootstrap \
  --asset vendor/icons/bootstrap-icons.svg --asset-kind svg --asset-name bootstrap-icons \
  --asset vendor/fonts/Inter-Variable.ttf --asset-kind font --asset-name inter \
  --emit-image output/pages_png \
  --emit-jit output/render.jit.jsonl \
  --emit-perf output/render.perf.jsonl \
  --out output/render.pdf

5. Use --repro-record / --repro-check once your layout stabilizes.

Practical tips:

• Compare against full-page exports when available.
• Keep a fixed preview DPI (for example 144 or 200) across iterations.
• Commit PNG baselines for repeatable visual checks.

Public Golden Regression Suite

Launch-grade render regression coverage is available under goldens/ with three fixtures:

• invoice
• statement
• menu

Golden contract assets:

• Expected hashes: goldens/expected/golden_suite.expected.json
• Expected PNG baselines: goldens/expected/png/<case>/<case>_page1.png

Run against committed expectations:

python goldens/run_golden_suite.py verify

Refresh baselines intentionally:

python goldens/run_golden_suite.py generate

Human + AI Operating Mode

Recommended automation defaults:

fullbleed --json-only render ...

Why this is agent-safe:

• For command-execution JSON payloads, schema is always present.
• Parser usage errors (exit=2) are emitted by argparse as usage text, not JSON payloads.
• ok indicates success/failure without parsing text.
• Optional artifacts are explicitly named in outputs.
• Schema introspection is available at runtime (--schema).

Example parse loop:

import json, subprocess

proc = subprocess.run(
    [
        "fullbleed", "--json-only", "render",
        "--html", "templates/report.html",
        "--css", "templates/report.css",
        "--out", "output/report.pdf",
    ],
    capture_output=True,
    text=True,
    check=False,
)

payload = json.loads(proc.stdout)
assert payload["schema"] == "fullbleed.render_result.v1"
assert payload["ok"] is True
print(payload["outputs"]["pdf"])

MACHINE_CONTRACT.v1

{
  "schema": "fullbleed.readme_contract.v1",
  "package": "fullbleed",
  "cli_entrypoint": "fullbleed",
  "dev_cli_entrypoint": "python -m fullbleed_cli.cli",
  "python_module": "fullbleed",
  "json_discriminator": "schema",
  "core_commands": [
    "render",
    "verify",
    "plan",
    "debug-perf",
    "debug-jit",
    "run",
    "finalize",
    "inspect",
    "compliance",
    "doctor",
    "capabilities",
    "assets",
    "cache",
    "init",
    "new"
  ],
  "result_schemas": [
    "fullbleed.render_result.v1",
    "fullbleed.verify_result.v1",
    "fullbleed.plan_result.v1",
    "fullbleed.run_result.v1",
    "fullbleed.inspect_pdf.v1",
    "fullbleed.inspect_pdf_batch.v1",
    "fullbleed.inspect_templates.v1",
    "fullbleed.compose_plan.v1",
    "fullbleed.compliance.v1",
    "fullbleed.capabilities.v1",
    "fullbleed.doctor.v1",
    "fullbleed.assets_list.v1",
    "fullbleed.assets_info.v1",
    "fullbleed.assets_install.v1",
    "fullbleed.assets_verify.v1",
    "fullbleed.assets_lock.v1",
    "fullbleed.cache_dir.v1",
    "fullbleed.cache_prune.v1",
    "fullbleed.init.v1",
    "fullbleed.new_template.v1",
    "fullbleed.new_list.v1",
    "fullbleed.new_search.v1",
    "fullbleed.new_remote.v1",
    "fullbleed.debug_perf.v1",
    "fullbleed.debug_jit.v1",
    "fullbleed.repro_record.v1",
    "fullbleed.error.v1"
  ],
  "artifact_flags": [
    "--emit-manifest",
    "--emit-jit",
    "--emit-perf",
    "--emit-glyph-report",
    "--emit-page-data",
    "--emit-compose-plan",
    "--emit-image",
    "--image-dpi",
    "--deterministic-hash",
    "--repro-record",
    "--repro-check"
  ],
  "fail_on": ["overflow", "missing-glyphs", "font-subst", "budget"],
  "budget_flags": ["--budget-max-pages", "--budget-max-bytes", "--budget-max-ms"],
  "profiles": ["dev", "preflight", "prod"],
  "pdf_version_default": "1.7",
  "known_exit_codes": {
    "0": "success",
    "1": "command-level validation/operational failure",
    "2": "argparse usage error",
    "3": "CLI runtime/input error wrapper"
  }
}

Important Behavior Notes

• render --json cannot be combined with --out - (stdout PDF bytes).
• verify defaults to stdout PDF unless --emit-pdf is provided; for machine mode, use --emit-pdf <path>.
• --emit-image <dir> writes per-page PNGs as <stem>_pageN.png (stem comes from --out/--emit-pdf, or render when streaming PDF to stdout).
• In template auto-compose runs, --emit-image artifacts are rasterized from finalized composed pages and report outputs.image_mode=composed_pdf; otherwise image_mode=overlay_document.
• outputs.deterministic_hash_mode is pdf_only by default and artifact_set_v1 when image artifacts are emitted.
• If both --emit-page-data and --emit-glyph-report are set, current engines use a combined API and render once; older engines without that API fall back to a double render.
• Production target is PDF 1.7.
• run accepts --html-str without requiring --html.
• init now scaffolds COMPLIANCE.md for project-level release review.
• compliance --json emits machine-readable legal/procurement diagnostics.
• --watermark-layer underlay is accepted as a legacy alias and normalized to background.
• --emit-manifest includes a watermark object with text|html|image|layer|semantics|opacity|rotation|enabled.
• --fail-on overflow is enforced from placement data and may auto-enable internal JIT planning.
• --fail-on font-subst is enforced using missing glyph and fallback diagnostics.
• --allow-fallbacks allows fallback diagnostics to remain informational for missing-glyphs / font-subst gates while still reporting them in JSON output.
• --fail-on budget requires at least one budget threshold flag.
• --repro-check fails on input/hash drift and lock hash mismatches when lock data is available.
• PDF/A and PDF/X/VT profiles require output intent metadata; pdfa4, pdfa4e, and pdfa4f emit PDF 2.0 automatically. PDF/A, PDF/X/VT, PDF/UA, and WTPDF text output enforces embedded-font constraints, and CLI errors include an actionable hint to add an embeddable font asset.
• --pdf-profile pdfua1 enables tagged output and PDF/UA-1 identification metadata. Treat verifier/seed traces as machine evidence before making external conformance claims.
• argparse usage errors exit with code 2 and emit usage text (not JSON), even when --json is present.

Related Docs

• Agent workflow guide: llm.txt
• CLI JSON contract quick reference: cli_schema.md
• CLI epoch/spec: CLI_EPOCH.md
• PDF template/XObject composition guide: docs/pdf-templates.md
• Licensing guide: LICENSING.md
• Third-party notices: THIRD_PARTY_LICENSES.md
• Living docs example project: examples/living_docs_atlas/README.md
• Roofing invoice parity example: examples/roofing_invoice/README.md
• Iconography smoke example: examples/iconography_test/README.md
• Public golden regression suite: goldens/README.md

License

Fullbleed is licensed under the MIT License (MIT). Commercial use, modification, distribution, and use in proprietary software are permitted subject to the notice-preservation requirement in LICENSE.

Cargo and PyPI metadata both declare MIT.

• Copyright notice: COPYRIGHT
• Third-party notices: THIRD_PARTY_LICENSES.md
• Practical licensing guide: LICENSING.md

For license information, please visit fullbleed.dev or email info@fullbleed.dev.

License integrity gate (CI-friendly, no build required):

python tools/check_license_integrity.py --json