Skip to main content
Pipelines are saved multi-step workflows built from nodes — remove background, upscale, generate, and more — chained together in a graph. Create pipelines in the Pixa web app, then run them from the CLI or MCP.

​
Running a pipeline

​
CLI

Use pixa run pipeline with the pipeline ID and inputs mapped to node names: 
pixa run pipeline pl_abc123 \
  --input photo=./product.png \
  --json
Inputs are name=value pairs where the name matches a node alias in the pipeline. Values can be local file paths (auto-uploaded), URLs, asset IDs, or text: 
pixa run pipeline pl_abc123 \
  --input photo=./product.png \
  --input prompt="white background" \
  --collection-id col_xyz789 \
  --json
FlagDescription
--inputNode input as name=value (repeatable)
--asyncReturn immediately with job ID
--dry-runValidate without executing
--collection-idSave results to a specific collection
JSON input: 
pixa run pipeline pl_abc123 --json-input '{
  "inputs": {
    "photo": "./product.png",
    "prompt": "white background"
  },
  "collection_id": "col_xyz789"
}' --json

​
MCP

Use the pipelines tool with action run: 
Run my "Product Photo Cleanup" pipeline on these three product images.
The agent calls pipelines (action: get) to inspect the pipeline, then pipelines (action: run) with inputs mapped to the correct node IDs.

​
Managing pipelines

​
CLI commands

SubcommandDescription
pipelines listList all pipelines (--limit, --after)
pipelines get <id>Get pipeline details including nodes and edges
pipelines createCreate a new pipeline (--name required)
pipelines update <id>Update a pipeline
pipelines search <query>Search pipelines by name
pipelines delete <id>Delete a pipeline
pipelines runs <id>List past runs of a pipeline
pipelines node-typesList available node types with I/O specs

​
MCP actions

The pipelines MCP tool supports these actions:
ActionDescription
listList all pipelines
searchFind pipelines by name
getGet pipeline details with nodes and edges
runStart a pipeline execution
list_runsList runs for a specific pipeline
get_runGet run details with per-node status
cancel_runStop a running execution
list_node_typesAvailable node types with I/O specs
Use display_pipeline to render an interactive node graph preview in MCP App UIs.

​
Finding pipelines

# List all
pixa pipelines list --json

# Search by name
pixa pipelines search "product photo" --json

# Get details by ID
pixa pipelines get pl_abc123 --json

​
Monitoring pipeline runs

With --async, pixa run pipeline returns a job ID immediately. Use pixa jobs follow to stream status updates: 
pixa jobs follow job_xyz789 --json
Set --interval to control the polling frequency (default is 2 seconds): pixa jobs follow job_xyz789 --interval 5s

​
End-to-end example

# 1. Find the pipeline
pixa pipelines search "product cleanup" --json

# 2. Inspect its inputs
pixa pipelines get pl_abc123 --json

# 3. Run the pipeline
pixa run pipeline pl_abc123 \
  --input photo=./product.png \
  --collection-id col_xyz789 \
  --json

# 4. Check results (if run with --async)
pixa jobs get job_def456 --json
For more on pipeline concepts and the web-based pipeline editor, see the Pipelines user guide.