Nodes
Every workflow in Pictell starts with a Node.
A node is the smallest functional unit/block in a visual workflow.
Each node does one clear job: it holds data and produces output.
By connecting nodes together, you create workflows.
What Is a Node?
A node represents a single step in your workflow.
Examples:
- Writing a prompt
- Uploading an image
- Running an AI model
- Displaying a result
Nodes are designed to be:
- Single-purpose – one responsibility per node
- Composable – easy to connect and reuse
- Predictable – the same input produces consistent behavior
Input and Output
All nodes participate in the same data flow model:
Input → (Optional Processing) → Output
However, not every node performs computation.
Data Nodes
Some nodes exist purely to provide or display data.
Examples:
- Prompt Node – stores and provides text input
- Reference Image Node – stores and provides image input
- Image/Video Result Node – displays generated output
- Start-end Frame Reference Image Node - stores and provides start frame and end frame image input for video generation
These nodes do not “run” models.
They pass structured data to other nodes.
Model Nodes
Model nodes are the core execution units in Pictell.
They are also the key nodes used to create automations.
A model node automatically retrieves data from all connected upstream nodes.
Based on its selected model and parameters, it generates a new result.
Unlike data nodes, model nodes perform actual computation.
They:
- Collect input from connected nodes
- Execute the selected AI model
- Produce output for downstream nodes
- Participate in workflow automation
When creating an Automation, only model nodes are selectable, because they define the executable steps of a workflow.
In Pictell, only model nodes are executable.
Node Classification
The following table lists all available nodes in Pictell, categorized as Data Nodes or Model Nodes:
| Node Name | Category | Description |
|---|---|---|
| Prompt | Data Node | Stores and provides text input for AI models |
| Reference Image | Data Node | Stores and provides image input (supports up to 4 images) |
| Image Result | Data Node | Displays generated image output |
| Video Result | Data Node | Displays generated video output |
| Start-End Reference Image | Data Node | Stores and provides start frame and end frame image input for video generation |
| Camera Control | Data Node | Provides camera control parameters (azimuth, elevation, distance) for 3D scene rendering |
| Image Model | Model Node | Executes image AI models (e.g., text-to-image, image-to-image) |
| Video Model | Model Node | Executes video AI models (e.g., image-to-video) |
Connecting Nodes
Nodes are connected by links that show how data moves.
- Left side — Input
- Right side — Output
Drag from the right side of one node to the left side of another.
If a node has no left handle, it cannot receive input.
If a connection cannot be created, the two nodes are not compatible.
When nodes are connected:
- Data flows automatically
- The order becomes clear
- The workflow becomes repeatable
Pictell is visual so you can easily see how everything connects.
Node State and Execution
Nodes store their current data and settings, such as prompts, images, model choices, and results.
Because of this, workflows can be run again with the same setup.
However, each run is fresh.
Every time you run a workflow:
- It uses the current inputs
- It generates new outputs
- It does not rely on hidden memory
This makes workflows simple, predictable, and safe to automate.
Why Pictell Is Node-Based
Using nodes allows you to:
- Build workflows step by step
- Experiment without breaking everything
- Reuse and recombine logic
- Scale from simple to complex workflows naturally
A single node is simple.
A connected set of nodes becomes powerful.
Next, we’ll look at how nodes work together as a Workflow,
and how execution is controlled through Autoflow and Phases.