Metaphora

Carry content across boundaries.

PiSrc Metaphora is a growing suite of AI-powered transformation tools that convert content from one form to another. PDFs become HTML. HTML becomes authored AEM components. English becomes any language, governed by your brand voice and editorial guidelines. Each product in the Metaphora family solves a specific content transformation problem that enterprise teams encounter daily, and each runs directly on your existing infrastructure without external processing servers. All Metaphora products require an OpenAI API key. Support for additional LLMs is in development.Metaphora products are available as standalone AEM enhancements or as part of the broader [Prism](/products/prism) AI services platform.

PDF to HTML

From static documents to living pages

PDFs are everywhere in the enterprise and they are a dead end for search engines, mobile devices, and accessibility tools. Metaphora PDF to HTML extracts the full content of a PDF and reconstructs it as clean, responsive HTML that is SEO friendly and portable across devices.

The conversion goes beyond text extraction. Metaphora analyzes the visual structure of each page, identifying background elements, matching fonts to web equivalents, preserving object positioning, and converting embedded images to optimized web formats. The output is not a flat text dump with a screenshot; it is a faithful HTML representation of the original document that respects layout, typography, and visual hierarchy.

The result is content that search engines can index, screen readers can parse, and visitors can read on any device. For organizations with large PDF libraries (product sheets, white papers, regulatory filings, annual reports), Metaphora PDF to HTML is the fastest path to making that content work on the web.

PDF to HTML

From static documents to living pages

AI Translator

Every language, your voice

Machine translation has come a long way, but it still produces output that reads like machine translation. The grammar is correct and the meaning is preserved, but the tone is flat, the terminology is generic, and the result needs a human editor to sound like it belongs on your website. Metaphora AI Translator closes that gap.

Metaphora AI Translator improves on conventional machine translation by incorporating context that standard engines ignore. It considers the surrounding text, the purpose of the page, your company's editorial guidelines, preferred terminology, and brand tone when producing each translation. A product page reads like a product page. A legal disclosure reads like a legal disclosure. Technical documentation preserves the precision of the original while adapting to the conventions of the target language. The difference between a mechanical word-for-word rendering and a translation that sounds native is exactly the kind of context that AI handles well.

A built-in glossary lets you define brand terms that should never be translated, or that should always be translated in a specific way. Product names, trademarks, and domain-specific terminology stay exactly as you intend them in every language. Translation memory stores previously approved translations so that recurring phrases and sentences are handled consistently across pages and updates. A translation cache reduces token usage by reusing results for content that has already been processed, keeping costs down as your translation volume grows.

Google Translate is available as a baseline translation option within Metaphora for teams that prefer it or want to compare output. OpenAI provides the advanced contextual translation layer. Both are accessible through the same interface. Metaphora AI Translator integrates directly into AEM as a translation project dashboard. Authors and localization managers kick off translation requests for individual pages or entire folder structures from within the AEM authoring environment. The dashboard tracks each request through its lifecycle and provides real-time budget spend estimates and cumulative usage reporting so teams can manage translation costs with full visibility.

Like PDF to AEM, the AI Translator runs directly on AEM Author servers. The only external dependency is the AI endpoint itself. No intermediary processing servers, no third-party translation management platforms, and no data routing through external APIs beyond the language model. This keeps the architecture simple, the data secure, and the operational overhead low.

AI Suite

How they work together

The Metaphora products are designed to compose. A regulatory PDF can be converted to HTML for immediate web publication, then transformed into authored AEM components for long-term content management, then translated into twelve languages with brand-consistent tone and terminology. Each step builds on the previous one, and all three run within your AEM environment.

For organizations using Prism, Metaphora-converted content is automatically available to Prism's unified search index. Translated pages are indexed in their respective locales. PDFs that were previously invisible to Prism's knowledge base become searchable, citable, and conversational the moment they are converted.

Tell Me More