A proprietary production workflow that transforms printed textbooks into interactive digital applications, compliant with the requirements of the Ministry of Education.
Turning printed textbooks and editorial PDFs into digital products involves a large volume of repetitive work: extracting and correcting text, identifying page structure, processing images, integrating multimedia resources, and final quality control.
During tender periods, these operations must be carried out at scale and in short timeframes, both for the DEMO version required for the submission, and for the final version to be delivered.
We built the complete technical workflow for producing digital textbooks, from processing PDFs to integrating interactive resources and generating the final packages.
Processing and structuring. Our proprietary platform analyzes native or scanned PDFs, identifies page elements, and transforms them into structured content. Operators review and correct every element in a visual interface.
Interactivity and compatibility. We turn textbooks into HTML applications that work offline, too, with interactive exercises, audio and video materials, animations, maps, and H5P content. Content is adapted for desktop, tablet, and phone.
Quality control. We maintain version history, run automated checks, and generate reproducible exports from a validated data source. We coordinate proofreading, testing, and acceptance for both DEMO and final versions.
The workflow includes asynchronous processing, locally running AI models, semantic search, and controlled access for AI agents, while maintaining human validation.
The production platform uses Python, FastAPI, React, and PostgreSQL, with asynchronous processes, models for document structure analysis, image processing, and locally run AI models via Ollama. Integration with AI agents uses RAG search and Model Context Protocol.
Textbooks are built with HTML, CSS, JavaScript, and H5P, and production automations use Node.js and Bash.
