askLLM – ask an LLM about your data inside jamovi
Posted: Tue Sep 15, 2026 8:39 am
Hello everyone,
I would like to share askLLM, a free and open-source module that lets you ask an LLM questions about your open dataset without leaving jamovi. You select the variables, type a question, and the answer appears in the results panel.
The current version (1.3.0) has two analyses under Analyses ▸ askLLM:
- jamovi Module Guider answers "which analysis should I run?" It recommends a jamovi analysis and quotes the exact menu path. The suggestions are grounded in the modules you actually have installed, so it does not invent menus that do not exist.
- R code tutor answers "how do I write the R code for this?" It gives you code to paste into the Rj Editor and run yourself. The code only uses packages that Rj actually bundles, so it will not suggest packages you do not have or invent file paths. Rj is desktop-only, so R code tutor needs desktop jamovi.
askLLM is a copilot, not an autopilot. It only advises: you run every analysis and every line of code yourself, and it never writes to your data. It also does not read jamovi's analysis output, so for multi-step work you summarise the previous result in your next question and the LLM advises from there.
The module works with NVIDIA NIM, Google Gemini, OpenRouter, GitHub Models (all with free tiers), a local Ollama model, or any OpenAI-compatible server. Only summary statistics of the selected variables and your question are sent, never the raw data rows. API keys are read from your local environment and are never written into the .omv file. With Ollama, nothing leaves your computer.
This release installs as a side-loaded .jmo on Windows 64-bit (tested with jamovi 28.2.0.0).
Project and installation information:
https://github.com/SCgeeker/askLLM
Download version 1.3.0:
https://github.com/SCgeeker/askLLM/releases/tag/v1.3.0
Choosing a model: https://scgeeker.github.io/askLLM/choose-model.html
Learn R with Rj: https://scgeeker.github.io/askLLM/learn-r.html
There is also a sister site, stat-skills-tutorials, on using an LLM as a copilot for statistics, with a strong focus on checking what the AI gives you (verification checklists, spot-the-error exercises, a prompt library):
https://scgeeker.github.io/stat-skills-tutorials/
Feedback, bug reports, and suggestions are very welcome, especially from anyone using it in teaching or research.
I would like to share askLLM, a free and open-source module that lets you ask an LLM questions about your open dataset without leaving jamovi. You select the variables, type a question, and the answer appears in the results panel.
The current version (1.3.0) has two analyses under Analyses ▸ askLLM:
- jamovi Module Guider answers "which analysis should I run?" It recommends a jamovi analysis and quotes the exact menu path. The suggestions are grounded in the modules you actually have installed, so it does not invent menus that do not exist.
- R code tutor answers "how do I write the R code for this?" It gives you code to paste into the Rj Editor and run yourself. The code only uses packages that Rj actually bundles, so it will not suggest packages you do not have or invent file paths. Rj is desktop-only, so R code tutor needs desktop jamovi.
askLLM is a copilot, not an autopilot. It only advises: you run every analysis and every line of code yourself, and it never writes to your data. It also does not read jamovi's analysis output, so for multi-step work you summarise the previous result in your next question and the LLM advises from there.
The module works with NVIDIA NIM, Google Gemini, OpenRouter, GitHub Models (all with free tiers), a local Ollama model, or any OpenAI-compatible server. Only summary statistics of the selected variables and your question are sent, never the raw data rows. API keys are read from your local environment and are never written into the .omv file. With Ollama, nothing leaves your computer.
This release installs as a side-loaded .jmo on Windows 64-bit (tested with jamovi 28.2.0.0).
Project and installation information:
https://github.com/SCgeeker/askLLM
Download version 1.3.0:
https://github.com/SCgeeker/askLLM/releases/tag/v1.3.0
Choosing a model: https://scgeeker.github.io/askLLM/choose-model.html
Learn R with Rj: https://scgeeker.github.io/askLLM/learn-r.html
There is also a sister site, stat-skills-tutorials, on using an LLM as a copilot for statistics, with a strong focus on checking what the AI gives you (verification checklists, spot-the-error exercises, a prompt library):
https://scgeeker.github.io/stat-skills-tutorials/
Feedback, bug reports, and suggestions are very welcome, especially from anyone using it in teaching or research.