Integrating LLM-Based Chatbot with H5P
Hi everyone,
I’m Kartavya Suhagiya, a student based in State College at PSU, working on a capstone project with my university. We’re developing a chatbot for pre-service teachers that simulates conversations with parents and provides feedback based on their responses. To achieve this, we plan to use LLMs like OpenAI and interact with a backend Python server via APIs.
I’m interested in working with the open-source version of H5P to create a new content type that allows JavaScript files to communicate with a backend server for real-time processing. Is there a recommended approach for integrating such API-based communication within H5P? Are there any examples or documentation that could help?
Additionally, I’m facing an issue when trying to upload an H5P file to my Moodle course. The system keeps rejecting the file, stating that a valid "content" folder is missing. I’ve checked multiple forums and verified that all required files and folder structures are in place, yet the issue persists. Could anyone provide insights into what might be causing this or how to debug and resolve it?
Any guidance on these topics would be greatly appreciated!
Best,
Kartavya
otacke
Tue, 02/11/2025 - 09:42
Permalink
a) There currently is no
a) There currently is no clean way to use external services in H5P content types, see https://h5p.org/node/1454313
b) Set the X flag and D flag when zipping, see https://h5p.org/comment/43244#comment-43244 - or don't do this manually to begin with but use the H5P CLI tool to build content types and pack libraries.