arcflow run

Execute a workflow file from the terminal. The CLI entry point exists for a uniform developer experience; current implementation directs you to the Python SDK for actual execution.

Usage

bash
arcflow run WORKFLOW_FILE [--input TEXT]
FlagDescription
WORKFLOW_FILEPath to .py or workflow script
--inputOptional run input string (workflow must read it)

Current behavior

bash
arcflow run workflows/example_workflow.py --input "hello"

Prints guidance to invoke the Python SDK:

text
[ArcFlow] arcflow run for workflows/example_workflow.py requires the Python SDK (arcflow package).
 Use: python -c "import runpy; runpy.run_path('workflows/example_workflow.py')"
[ArcFlow] --input is accepted; wire your workflow entrypoint to call workflow.run().

Exit code 2 (execution not completed via CLI native path).

Python:

python
from arcflow import Agent, Workflow

wf = Workflow("demo")
wf.step(Agent(name="writer", role="author", instructions="Summarize."))
result = wf.run("hello")
print(result.output)

Or run the workflow file directly:

bash
python workflows/example_workflow.py

TypeScript: use @arcflow/sdk with node or tsx per quickstart-typescript.md.

Remote runtime (--runtime planned)

Future CLI versions may include --runtime http://host:8080 for server-backed runs. When wired, pattern will be:

bash
export ARCFLOW_SERVER_API_KEY=dev-secret
arcflow run workflows/demo.py --runtime http://localhost:8080 --input "hello"

Until native remote run ships, use curl POST /v1/runs or Python Workflow(..., runtime="http://localhost:8080").

Exit codes

CodeMeaning
0Success (when native execution implemented)
2Execution failure or SDK delegation message
3Infrastructure / parse error