Core concepts
The objects VOLT works with, from teams and clusters to pipelines, analyses and exposures, and how they relate.
Most of VOLT reads naturally once you know a dozen words. This page defines them and shows how they connect.
Workspace
Team. The workspace everything belongs to: trajectories, clusters, plugins, notebooks, whiteboards, AI conversations, secret keys and roles. Switching teams swaps the whole workspace. Members get a role, and the role's permissions decide what they can see and change. See Teams.
Cluster. A machine enrolled into a team that stores the team's data and runs its work: trajectory processing, analyses and notebooks. A team needs at least one connected cluster. In the browser and the desktop app the cluster is your own tab or window; on a server you enroll workstations, servers or HPC nodes. See Clusters.
Deployment mode. A server runs in cloud mode (accounts, sign-up, teams, invitations) or in local mode (one local account, signed in automatically), which is what the browser and the desktop app use. See Configuration.
Data
Trajectory. One simulation as VOLT keeps it: the files you uploaded, split into frames, with the metadata, previews and 3D models derived from them. See Trajectories.
Timestep. One frame of a trajectory, identified by the simulation timestep it was written at. Analyses run per timestep, and you choose which timesteps a pipeline covers.
Simulation cell. The box of a frame: its cell vectors, origin and periodic boundary conditions. VOLT extracts it from every frame. See Simulation cells.
Analysis
Plugin. A packaged analysis algorithm, published to the plugin registry and installed into a team. A plugin is a workflow: its arguments, the program it runs on each frame, and the results it exposes. See the plugin catalog and Writing plugins.
Pipeline. An ordered list of stages you build in the canvas. Its modifications are the stages: plugin analyses, slice planes and expression selections, which the cluster applies in order so each one sees what the previous left, plus color coding, which only changes what you see. Its visual elements decide what is drawn. See Pipelines & analyses.
Pipeline run. One execution of a pipeline over the timesteps you picked. A run keeps its stages and timesteps, so you can reopen it, extend it with new stages, or run it again with changed parameters.
Analysis. The result of one plugin stage of a run: the plugin, its arguments, its status and per-frame progress, and its outputs. A stage that matches one already computed reuses that analysis instead of computing it again.
Exposure. One named output of an analysis, such as a structure classification or a set of dislocation lines. It can hold tables, per-atom properties and 3D models, stored on the cluster as Parquet tables and compressed GLB models, and read back as listings, columns of the atoms or layers in the scene. See Analysis results.
Job. One unit of work on a cluster's queue, such as processing one frame of a trajectory or running one frame of an analysis. See Jobs.
Programmatic access
Secret key. A team credential for programs: the SDKs, scripts and notebooks. A key carries a role, so it can do exactly what that role allows. See Secret keys.
Notebook. A Python notebook whose kernel runs on one of the team's clusters (in the page, for the browser and desktop runtimes), with the voltsdk client already authenticated. A notebook is general, or linked to one trajectory. See Notebooks.