Assistant

Local language model

A chat panel in Studio that edits your workspace and answers questions about your data. The model runs on your computer.

Studio includes a language model that runs locally. It opens as a chat panel next to the plot. You can ask it to rename a set of scans, change the colour of a trace, or explain what a voltammogram or Nyquist plot shows.

The model is downloaded once, on first use, and then works without an internet connection. Your data is not sent anywhere.

When to use it

Renaming or tagging many traces at once, for example by the analyte concentration recorded in their metadata.
Plot edits in plain language, such as making one curve dashed or changing a colour.
A first reading of a result, such as what a Nyquist plot suggests for Rct and Cdl.

How it works

The model has access to the scans, parameters, and fit results in the current workspace and to the plot controls. When you ask for a change, it makes the same edit you would make by clicking, and the result appears in the plot and the experiment list.

Because the model runs on the CPU or GPU of your own computer, response time depends on your hardware.

Parameters

Model

Downloaded on first use and stored locally. No account or API key is needed.

Panel

Opened from the right rail of the main window. Type a request and press Enter.

Example

Rename a CV calibration series

Open a folder of CV scans recorded across a concentration sweep.
In the chat panel, type: "Rename each scan with the concentration from its metadata."
The scan list updates with the new names.
Ask: "Change the colour of the 50 nM trace to red."

Result. A labelled, recoloured scan list, ready for peak finding and a calibration curve in the Publication Workbench.

In Sensitify Studio

The chat panel is opened from the right rail of Sensitify Studio. It is included with the software; the model downloads on first use.