arcflow init

Scaffolds a new ArcFlow project directory with workflow folders, example stub file, and minimal config. Works offline (no network calls).

Usage

bash
arcflow init [OUTPUT_DIR] [options]
Argument / flagDefaultDescription
OUTPUT_DIRmy-arcflow-projectTarget directory
--langpythonpython or typescript (file extension)
--forceoffOverwrite non-empty directory

Examples

Python project:

bash
arcflow init my-bot --lang python

TypeScript project:

bash
arcflow init my-bot --lang typescript

Force into existing folder:

bash
arcflow init./existing --force

Generated structure

text
my-arcflow-project/
 arcflow.config.yaml
 workflows/
 example_workflow.py # or.ts
 agents/
 tools/

arcflow.config.yaml (initial content):

yaml
runtime_mode: embedded

workflows/example_workflow.{py|ts} contains a one-line comment pointing to arcflow run.

Expected output

text
[ArcFlow] Created project at my-arcflow-project. Next: cd my-arcflow-project && arcflow run workflows/example_workflow.py

Exit codes

CodeCause
0Success
1Directory not empty without --force
3Filesystem error creating dirs or files

Next steps after init

  1. Replace example workflow with real Workflow + Agent definitions (Python or TypeScript SDK).
  2. Copy .env.example patterns from SDK README for provider keys.
  3. Run via SDK or arcflow run (see run.md).

init does not install Python packages or npm modules. Install SDK separately per getting-started/install-and-build.md.