Create AI assistants for education integrated in your Learning Management System #

Create AI assistants for education integrated in your Learning Management System

LAMB is a web platform that allows you to design, train and publish AI-based learning assistants in a visual and intuitive way. It works as a “teaching chatbot builder” that combines language models (GPT-4, Mistral, local models) with your own educational materials.

LAMB is an open source project developed by Marc Alier and Juanan Pereira, professors and researchers at the Universitat Politècnica de Catalunya (UPC) and Universidad del País Vasco (UPV/EHU).

What does LAMB do for you? #

🎓 Specialized Subject Tutors #

Design assistants that only respond about the subject you choose, always staying within the appropriate educational context.

📚 Intelligent Knowledge Ingestion #

Upload documents (PDF, Word, Markdown) and LAMB processes them automatically with a flexible data model that:

• Extracts and structures content while preserving context and relationships
• Creates semantic embeddings optimized for educational search
• Allows custom metadata for each document
• Adapts to different formats and content structures
• Feeds the model through RAG (Retrieval Augmented Generation)

🔍 Advanced Testing and Debugging #

“Debug” mode that shows the complete prompt to understand exactly what is sent to the model, facilitating response optimization.

🎯 LTI Integration with Moodle #

Publish the assistant as an external LTI tool and embed it in your Moodle in a couple of clicks.

🔒 Guaranteed Privacy #

Students interact within LAMB; their data is not shared with external AI model providers.

Who is LAMB for? #

• 📖 Teachers and trainers who need a virtual assistant focused on their specific curriculum
• 🏫 Educational centers that use Moodle or other LMS and need to integrate AI without exposing student data
• 💡 Innovation teams experimenting with different LLMs and need a unified management panel

Main Features #

Unlimited Assistants #

Each with their own instructions, tone and personalized limits.

Flexible Knowledge Bases #

• Adaptable data model: flexible architecture that allows different types of content and structures
• Support for PDF, DOCX, Markdown (more formats coming soon)
• Public or private bases according to needs
• Vector embeddings system for semantic search
• Connectors in development for external sources (Google Drive, YouTube, APIs)

Multiple AI Models #

• OpenAI GPT-4o
• Mistral
• Local models
• One-click model switching

Automatic Citations #

The assistant provides responses with references to the source documents used.

Total Portability #

• Export/import assistants in JSON format
• Easy versioning and sharing

Multilingual Interface #

Catalan, Spanish, English and Basque included as standard.

Robust Access Control #

• Secret keys for registration
• Private bases to prevent unauthorized use

Growing Ecosystem #

• Modular and extensible architecture: designed to incorporate new features without affecting the core
• Customizable ingestion plugins for different data sources
• Open API for third-party integrations
• Continuous updates without dependence on a single AI provider
• Flexible data model that evolves with educational needs

In short #

LAMB gives you total control to build a “specialized ChatGPT” for your subject, connect it to your Moodle and keep your students’ data completely secure.

🎓 Research & Academic Foundation #

LAMB is built on solid academic research and adheres to the Safe AI in Education Manifesto - a comprehensive framework for ethical, secure, and educationally-aligned AI deployment.

📚 Academic Publication #

If you use LAMB in your research, please cite our work:

“LAMB: An open-source software framework to create artificial intelligence assistants deployed and integrated into learning management systems”

• Authors: Marc Alier, Juanan Pereira, Francisco José García-Peñalvo, Maria Jose Casañ, Jose Cabré
• Journal: Computer Standards & Interfaces, Volume 92, March 2025
• DOI: 10.1016/j.csi.2024.103940

🏛️ Academic Partners #

• Universidad del País Vasco (UPV/EHU) - Research institution and development partner
• Universitat Politècnica de Catalunya (UPC) - Research institution and development partner
  • Barcelona School of Informatics
  • Institut de Ciències de l’Educació - ICE
  • Department of Service and Information System Engineering (ESSI)

🙏 Acknowledgments #

Special thanks to the Open WebUI Project, Tsugi Project (Dr. Chuck Severance), TEEM Conference community, and Tknika Basque VET Applied Research Centre for their support and collaboration.