local-usage.md

Running Locally (without Docker)

Note

The primary and recommended way to use this tool is via Docker Compose, as described in the main README.
This guide is for cases where you are running the Dify Plugin Daemon locally (not as a container) and want to run the packager on the same machine.

• ✅ Prerequisites
• 📝 Quick Start
• ⚙️ Configuration

✅ Prerequisites

• uv available in your PATH
• Internet access (to download plugins and Python packages)

Important

Wheels are downloaded for the CPU architecture of the machine running this tool. Make sure you run it on the same architecture (amd64 / arm64) as your Dify Plugin Daemon.

📝 Quick Start

Clone the repository and run the script directly from the repository root with uv run:

git clone https://github.com/kurokobo/dify-plugin-offline-packager.git
cd dify-plugin-offline-packager

# From Dify Marketplace
uv run scripts/packager.py --marketplace "langgenius/openai:0.3.2"

# From GitHub Releases
uv run scripts/packager.py --github "junjiem/dify-plugin-agent-mcp_sse:0.2.4:agent-mcp_sse.difypkg"

# From a local file  (place it in ./difypkg/ first)
uv run scripts/packager.py --local "difypkg/my-plugin.difypkg"

The script resolves all paths relative to the current working directory, so always run it from the repository root.

Both the original and the offline-packaged file are saved to ./difypkg/:

difypkg/
  langgenius-openai_0.3.2.difypkg            ← original (online)
  langgenius-openai_0.3.2-offline.difypkg     ← offline-ready

⚙️ Configuration

Settings can be provided as environment variables (same as the Docker-based workflow):

Variable	Default	Description
DIFY_PLUGIN_DAEMON_VERSION	0.5.3	Version of the dify-plugin CLI binary to download. Should match your locally running Plugin Daemon.
MARKETPLACE_API_URL	https://marketplace.dify.ai	Dify Marketplace API URL
GITHUB_API_URL	https://github.com	GitHub URL (set for GitHub Enterprise)
PIP_INDEX_URL	https://pypi.org/simple	PyPI mirror URL