Skip to main content

Rivet CLI - run Command

Run a Rivet graph in a project using provided input values.

Quick Start​

# Run with basic input
npx @ironclad/rivet-cli run my-project.rivet-project --input name=Alice

# Run with JSON input
echo '{"name": "Alice"}' | npx @ironclad/rivet-cli run my-project.rivet-project --inputs-stdin

# Run specific graph
npx @ironclad/rivet-cli run my-project.rivet-project "My Graph" --input name=Alice

Description​

The run command executes a Rivet graph with specified inputs. This is particularly useful for:

  • Testing graphs with specific inputs
  • Integrating Rivet into command-line scripts and tools
  • Automating graph execution from other programming languages
  • Development and debugging of graph implementations

Usage​

The basic usage will run the main graph in the provided project file, with no input values:

npx @ironclad/rivet-cli run my-project.rivet-project

You can also specify a specific graph in the file to run:

npx @ironclad/rivet-cli run my-project.rivet-project "My Graph"

Inputs​

Inputs can be provided in two ways. The first way is to use the --input flag for each input value:

npx @ironclad/rivet-cli run my-project.rivet-project --input input1=5 --input input2=10

This is useful for basic input values and allows for easy testing of various scenarios.

The second way is to provide the inputs using a JSON object from standard input. This is useful for more complex input values:

echo '{"input1": 5, "input2": 10}' | npx @ironclad/rivet-cli run my-project.rivet-project --inputs-stdin

This is useful for more complex input values, such as arrays or objects, as well as piping input values from other commands or scripts.

Outputs​

The Rivet CLI outputs JSON data to standard output. Each Graph Output node in the graph will correspond to a key in the output JSON object.

The value of each property will be a Data Value object, with a type property and a value property.

For example, if a graph has two Graph Output Nodes, output1 (a string) and output2 (a number), the output JSON object will look like this:

{
  "output1": {
    "type": "string",
    "value": "Hello, World!"
  }
  "output2": {
    "type": "number",
    "value": 42
  }
}

Options​

Input Configuration​

  • --input - Specify an input value for the graph. Can be used multiple times.
  • --inputs-stdin - Read input values from standard input as a JSON object. Overrides any input values provided with --input.
  • --context - Specify a single context value for the graph. Can be used multiple times. Context can be used to pass global values to the graph. Context is specified using the same format as --input.

Output Configuration​

  • --include-cost - Includes the cost of the graph execution in the output JSON object. The cost is included as a cost property on the output JSON.