Our Technology

Discover some of the innovative libraries and APIs powering our AI-driven regulatory intelligence solutions.

Our Libraries

While maintaining our boutique approach of carefully cherry-picking the best proprietary and open source technologies available, we've developed our core components to better support our AI-driven solutions. These specialized libraries have evolved to such a degree that we're able to share them with the wider community. Some are freely available as open-source projects, while others can be accessed upon request.

Marmoset Document Processing API

A sophisticated document intelligence API powered by large language models that seamlessly transforms content from both URLs and uploaded files into clean, structured Markdown. The system specializes in extracting organized data from complex official sources, including EU regulatory documents and committee websites, making unstructured information accessible and usable.

Key Features:

Supports processing of HTML/Web pages, PDFs, Word documents, plain text, and Markdown files
Specialized content processing templates for different document types
Synchronous and asynchronous processing options
WebSocket support for real-time progress updates

Available on RequestRequest Access

Speech92

A high-performance text-to-speech conversion system built with FastAPI, which transforms written content into natural-sounding speech using advanced, local, neural models. The system handles document processing and asynchronous audio generation. Testable on the Wave92 website (use coupon WAVE92FRIENDS).

Key Features:

GPU-accelerated text-to-speech processing with multi-GPU queue management
Automatic document parsing for PDF and Word files with intelligent formatting preservation
Markdown support with special handling for headings, lists, and formatting
Asynchronous processing with real-time status updates and email notifications
Supabase integration for secure data storage and user history

Commercial API with free tier (first 5 minutes free)Try it on Wave92.com

VoiceStreamAI

An open-source Python-based server and JavaScript client solution that enables near-realtime audio streaming and transcription using WebSocket.

Key Features:

Real-time audio streaming through WebSocket
Modular design for easy integration of different VAD and ASR technologies
Support for multilingual transcription
Customizable audio chunk processing strategies

Open Source LibraryGithub

Interested in Our Products and Technology?

Subscribe to our Substack Newsletter to stay informed about our next releases.

Stay Informed