Skip to main content

SKILLJAR: AI chatbot widget for Skilljar courses

  • January 20, 2026
  • 3 replies
  • 185 views

Hi all -

I spent a few hours this weekend creating a project that some of you might find useful. I’ve been looking for an in-course AI assistant that embeds directly into your Skilljar LMS courses. As it turns out, the only easy option for Rise courses was via an iframe which wasn’t what I was after.

What I built is a floating in-course AI chat widget that allows learners to ask questions about your course. You can upload your course content via the tool (transcripts, lesson text, etc.) and the LLM (Claude, GPT, Gemini) will respond with that context in mind.

Here’s the general idea behind how it works:
• Select your LLM and enter your API key (Claude, GPT, Gemini).
• Upload your course materials (lesson text, videos, transcripts, PDFs).
• Set up the widget (title, theme, placement, behavior, guardrails).
• Add a single JavaScript snippet to the header of your Skilljar theme.
• Scope it to the right e-learning courses - so the chatbot only appears on the courses you want.

You can get the code here: https://github.com/kelly-vibe-coding/in-course-assistant

Hit me up with any questions.

3 replies

revathimenon
Forum|alt.badge.img+8
  • Gainsight Community Manager
  • January 20, 2026

Thanks for sharing this with the community ​@kellymullaney 😃🙌🏼


soniamoaiery
Forum|alt.badge.img
  • Gainsight Employee ⭐️
  • January 27, 2026

@kellymullaney Just sharing our latest open beta release today with you here! Thought you might be interested! 

 

 


  • Author
  • Contributor ⭐️
  • January 27, 2026

Thanks for sharing this, Sonia. It's probably cool to see someone in the community building something similar, as it validates the direction you were already heading. Kudos to your team for shipping what people actually want.

I would love to see SCORM support, but I completely understand that it's a bit more challenging. If you're looking to eventually support SCORM, I'd think most users would be happy with simply uploading the training data and then using the prompt to provide context to the LLM. In some ways, that's probably a better model anyway, since you can upload details that may not be contained in the course itself--like knowledge base articles, etc.