Docs
What is a plix?
A plix is an executable presentation: it holds the data, the Python cells that generate its slides, and their pinned environments.
Who is a plix for?
A plix is for individuals and teams who want portable, traceable and reproducible pipelines, presented where they are computed. Data, code, environments and figures travel together: a single source of truth reproducing the deck.
What can I put on a slide?
Traditional content such as text, images and tables, along with LaTeX equations, Mermaid diagrams, interactive plots such as Plotly figures, and native scientific views: molecular structures (PDB, XYZ), 3D models, FEM meshes and IFC building models.
What is the computational backend?
Code is organized as a Directed Acyclic Graph (DAG) of typed Python cells, run in dependency order.
Is caching implemented?
Yes. A cell whose code, inputs and environment are unchanged is not recomputed, so a Run only executes the sub-DAG downstream of a change.
What data types can code consume?
Every port of a cell has a kind. A kind defines the Python object the cell receives or returns and the file format in which the data is stored or uploaded.
| Kind | File | Python |
|---|---|---|
| Images | png | PIL.Image |
| Plotly | json | plotly.graph_objects.Figure |
| Box | yaml | int | float | str | list | dict |
| Table | csv | dict[str, numpy.ndarray] |
| Dataset | h5 | numpy.ndarray | dict[str, numpy.ndarray] |
| 3D model | glb | trimesh.Scene |
| FEM mesh | h5 | trimesh.Trimesh |
| IFC model | ifc | ifcopenshell.file |
| Atoms | pdb | Bio.PDB.Structure.Structure |
| Atoms | xyz | ase.Atoms |
Text, equations, diagrams and shapes are typed on the slide; a cell never reads or writes them.
Where does the cell run?
By default, the cells run in the Browser through Pyodide (WebAssembly). For heavy computation, a cell can run in the cloud using either a CPU or a GPU backend.
| Backend | CPU | Memory | GPU | Max time per cell | Compute hours per hour |
|---|---|---|---|---|---|
| Cloud-CPU | up to 2 cores | up to 4 GB | none | 10 min | 1 per core-hour used; idle time counts as 1 core |
| Cloud-GPU | up to 2 cores | up to 4 GB | 1 NVIDIA L4 (24 GB) | 10 min | 14.4 per hour for the GPU, plus the CPU's |
What is a compute hour?
1 compute hour = 1 CPU core for 1 hour. Resets monthly, no rollover.
Can I add my own packages?
Yes. Each cell runs in an environment where you list pip packages or public GitHub repositories that are Python packages. GitHub packages need a Cloud backend.
Can I add a private GitHub repository to a cell's environment?
No, at the moment only public GitHub repositories can be added.
How do I create a plix with AI?
Can I still edit my plix manually?
Yes, any aspect of a plix, including coding and environment management, can be edited manually.
How do I share a plix?
Turn sharing on to get a link. Anyone who opens it signs in and sees the plix read-only; Import copies it, with its data, code and environments, into their own account. Turning sharing off or generating a new link revokes every viewer. Only the plix's creator or a team admin can share it.
Questions: contact@plixlab.com