---
title: "Open Source Attributions"
description: "Open-source components Kliper builds on, and the licences they are used under."
version: "en"
---

> Documentation Index
> Fetch the complete documentation index at: https://docs.kliper.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Open Source Attributions

Kliper is built on open-source software. This page credits the components doing significant work inside the platform and records the licence each is used under.

These are **not** [subprocessors](/subprocessors). Every library listed here runs inside Kliper's own infrastructure — no customer data is sent to the projects or the people who wrote them. Services that do receive customer data are listed on the subprocessors page instead.

Last updated: **2026-08-03**.

## Document processing

- [pdf-inspector](https://github.com/firecrawl/pdf-inspector) — **By [Firecrawl](https://github.com/firecrawl) · MIT Licence · [github.com/firecrawl/pdf-inspector](https://github.com/firecrawl/pdf-inspector)** Converts text-based PDFs into structured Markdown. Kliper uses it on uploaded **client evidence**, where it keeps tables intact rather than flattening them into disconnected words, drops repeated page headers and footers, and reports when a PDF is a scan with no text layer instead of returning an empty result. Written in Rust and compiled to a native module, it runs entirely on our own servers — documents are never sent anywhere to be converted. Our thanks to the Firecrawl team for releasing it under a permissive licence.

- **pdf-parse** — **Apache-2.0 Licence** Extracts plain text from PDFs. Kliper uses it for **compliance documents** — ROC and AOC templates and the PCI DSS standard itself. Those are forms, and their control numbering survives plain text extraction more faithfully than structured conversion, so the two extractors are used deliberately for different jobs.

- **mammoth** — **BSD-2-Clause Licence** Extracts text from Word documents (`.docx`, `.doc`), used wherever evidence or a report arrives as a Word file.

## Security

- [ClamAV](https://www.clamav.net) — **GPL-2.0 Licence · [clamav.net](https://www.clamav.net)** Open-source antivirus engine with regularly updated definitions. Every uploaded file is scanned before it is stored. See [Security & AI Trust](/security-and-ai-trust).

## Platform

- **PostgreSQL + pgvector** — **PostgreSQL Licence** The database behind Kliper, with `pgvector` providing the vector similarity search that powers Cortex's knowledge retrieval.

- **Next.js, React, Tailwind CSS** — **MIT Licence** The web application, its component model, and its styling system.

- **shadcn/ui + Radix UI** — **MIT Licence** The interface component library, providing accessible primitives — dialogs, menus, tooltips — that keyboard and screen-reader users depend on.

- **Prisma** — **Apache-2.0 Licence** The database toolkit used for schema management and queries.

## A note on licences

Permissive licences such as MIT, Apache-2.0, and BSD allow commercial use provided the original copyright and licence notices are preserved. We keep those notices intact in the distributed packages and credit the projects here.

If you maintain a project listed on this page and would like the attribution corrected or worded differently, write to [security@kliper.dev](mailto:security@kliper.dev) and we will update it.

Source: https://docs.kliper.dev/open-source/index.mdx