# Known limitations — PDFluent SDK

Source: https://pdfluent.com/docs/known-limitations/

Support

# Known limitations

Generally available Updated 2026-05-14

[Star on GitHub](https://github.com/pdfluent/pdfluent-sdk)

### On this page

- Release status
- Structured extraction
- Rendering and XFA
- PDF/A scope
- Platform matrix
- Performance claims
- Not Adobe parity

## Release status

The PDFluent SDK 1.5.0 is available for Python, Rust, Node.js, WebAssembly,.NET, Java. Pin the version you ship and read the changelog before upgrading. The desktop editor is generally available since 1.0.0 (2026-09-18) [B45] and is versioned separately from the SDK.

## Structured extraction

- Structured extraction ships as Preview: output quality is still improving and the API may change in a minor release. Plain-text extraction is not Preview.
- The `balanced` tier is native only and never WebAssembly. It runs on glibc Linux, macOS and Windows; on musl Linux loading returns a typed `ModelError`.
- Models are a separate, explicit install: `pdfluent models install layout tables`. The `pdfluent` command: from SDK 1.3.0, the Python package from PyPI installs it [B99]; in other setups it comes from the CLI archives on the release page. Nothing downloads automatically at run time. Without the installed models (or, in Rust, the cargo features), the tier refuses with `MODEL_NOT_AVAILABLE`.
- Plain-text extraction can be slightly slower on some documents: combining marks are now placed after the glyph they mark, which adds a small per-glyph cost. Follow-up work is planned after 1.1.0.
- Node.js on Linux (glibc): the `@pdfluent/node` addons need at most GLIBC_2.17, so glibc 2.17 or newer works. The pdfluent CLI's linux-gnu binaries need glibc 2.18 or newer. [B50, 2026-10-16]

## Rendering and XFA

Flatten of XFA forms to static PDFs is reliable and is the recommended production path. Live rendering of complex dynamic XFA forms can produce visual deltas against Adobe Reader; the golden-image set we measure against is improving wave over wave.

## PDF/A scope

The PDF/A converter and validator cover PDF/A-1b, PDF/A-2b, and PDF/A-3b. Other PDF/A profiles (1a, 2a, 2u, 3a, 3u, 4 and its family) are scoped for later releases.

## Platform matrix

Pre-built binaries cover the common platforms in each channel. Less common architectures or runtimes may need a build-from-source step — check the channel page for [Rust](https://pdfluent.com/docs/rust/), [Python](https://pdfluent.com/docs/python/), [WASM / JS](https://pdfluent.com/docs/wasm/), [.NET](https://pdfluent.com/docs/dotnet/), [Java](https://pdfluent.com/docs/java/), or [Node.js](https://pdfluent.com/docs/node/).

## Performance claims

Every performance figure on the site carries a claim ID and a measurement date. Figures without one were withdrawn. A figure holds for the machine and build it names.

## Not Adobe parity

PDFluent is not a drop-in Adobe Reader replacement. We do not claim full Adobe rendering parity. We aim for reliable flatten, reliable text extraction, reliable PDF/A conversion, and improving render fidelity wave over wave.
