Financial education that listens, explains, and adapts.

Challenge: Financial education doesn’t fail because of its content. It fails because of its format.

Many learners do not struggle with complex financial topics because the content is too difficult. They struggle with formats that do not allow for questions, do not offer individual customization, and cannot incorporate their own materials. Traditional learning offerings are too rigid for a subject that depends so heavily on personal life situations and prior knowledge.
At the same time, educational and research contexts impose specific requirements: systems must be transparent, auditable, and GDPR-compliant. Proprietary platforms are therefore often ruled out.

Solution: An AI platform that does more than just provide answers

As a subcontractor in the MINTIFY project, VAGO Solutions is developing the complete AI solution: a modular platform that combines chat, voice interaction, RAG, OCR, multiple learning modes, and a structured assessment system into a single application.
Learners can interact with the system via text or voice. The platform supports text-to-text, speech-to-text, and text-to-speech, as well as full-fledged voice dialogs via an integrated STT-LLM-TTS pipeline. This allows financial topics to be not only read but also explored through dialogue in an accessible manner.
A central element is the RAG component: Users can upload their own files—such as PDFs, Office documents, or images—and use them as a personal knowledge base. Scanned documents are also processed via OCR. The system thus answers learning questions not only based on general model knowledge but also on the basis of individually contributed materials.

Technical Highlights: Didactic Depth and Technical Control

The platform operates in several didactic modes. In teaching mode, the AI explains content step by step, asks comprehension questions, and provides exercises. In “Learning Buddy” mode, it acts as a motivating coach who asks follow-up questions and provides summaries. These modes are supplemented by an online search function with source citations and a simulated agent shopping mode. For quality assurance, the platform integrates a judgment panel: Three separate, color-coded evaluation fields—each generated by its own model run—make answers verifiable from both a didactic and technical perspective. All project artifacts, including source code, configurations, prompt sets, and documentation, are publicly released under the MIT or Apache 2.0 license.

• Multimodal interaction via text and speech
• RAG with OCR support for user-provided files
• Integrated judgment panel for quality assurance

Result: A reusable AI platform for education and research.

MINTIFY creates a transparent, autonomous, and pedagogically sound AI solution for financial education that addresses both research requirements and practical learning scenarios. The combination of interactive speech use, an individualized knowledge base, and a verifiable assessment system sets this solution far apart from traditional digital learning formats. And because everything is open source, it has the potential to become more than just a research project.