OpenAction
OpenAction is an AI-native action framework for the Agentic Home.
Move beyond rigid "If-This-Then-That" rules and tedious UI-based configurations. Powered by AI agent technology, OpenAction replaces traditional smart home solutions with flexible, adaptive logic. It helps to translate natural language intent into dynamic, executable scripts — giving your smart home the ability to adapt, reason, and act.
Read the full architectural deep-dive on Medium: From Smart Home to Agentic Home
Core Features
- Natural Language Automation: Define complex home behaviors via chat. No more wrestling with nested "If-Then" menus or YAML configurations.
- Stateful Intelligence: An integrated
Storeprovides persistence, allowing scripts to save and retrieve data across multiple executions. - Dynamic Scheduling: A built-in Cron service triggers AI-generated scripts based on time, sensor events, or external API data.
- Python Execution: The AI generates standard Python code, enabling complex calculations, loops, and sophisticated error handling.
Architecture & Workflow
The system acts as a bridge between high-level reasoning (Agents) and low-level hardware control. Here, the AI agent acts as a software developer who can introspect services and write custom, deterministic Python scripts to control smart home devices. The workflow is as follows:
- The Intent: The user describes a goal to an MCP-capable client (e.g., Claude Desktop): "If it’s dark outside and someone is home, turn on the living room light. If we’re away, simulate presence by randomly toggling lights, ending no later than late in the evening."
- The Translation: The Agent uses the OpenAction MCP Server tools to inspect available devices and writes a custom Python script.
- The Registration: OpenAction stores the script and sets up the necessary triggers (e.g., polling a weather API or listening for a sensor change).
- The Execution: When triggered, the script runs in a local sandbox, calling the endpoints of connected Sensors & Actuators.
How to Use It
1. Running via Docker
You can run OpenAction as a Docker container. Use host networking to allow the server to easily discover and communicate with local network devices. Mount a local directory to preserve the working state (scripts and store data) across container restarts.
sudo docker run -d --name openaction --restart always --log-driver local --log-opt max-size=10m --log-opt max-file=3 --network host -v /etc/script/openaction:/etc/work -e devices='' grro/openaction:0.0.30
By default the OpenAction server is running on port 8345. Make sure to map a persistent volume to /etc/work so that your environment state and custom scripts are not lost when the container is recreated.
2. Testing with MCP Inspector
You can test the server capabilities and available tools of the running Docker container using the official MCP Inspector via Server-Sent Events (SSE):
npx @modelcontextprotocol/inspector sse http://localhost:8345/sse
3. Integrate into an Agent
OpenAction works with any agent that supports MCP.
Claude Desktop example
To integrate OpenAction with Claude Desktop, add the MCP Server URL to the "mcpServers" section in your claude_desktop_config.json.
You may use mcp-proxy to bridge the Streamable HTTP transport and the stdio MCP transport used by Claude Desktop.
{
"mcpServers": {
"openaction": {
"command": "mcp-proxy",
"args": [
"http://localhost:8345/sse"
]
}
}
}
Use Claude Desktop
After setting up the MCP Server and integrating it with Claude Desktop, you can start describing your desired automations in natural language. The Agent will be able to generate and modify the task scripts and manage their execution based on your defined triggers.
Explore OpenAction Tools
Here is an example showing how you can ask Claude to list all available OpenAction tools.
View Registered Tasks
You can easily inquire about currently registered tasks and their statuses.
Create a New Task
Simply instruct Claude to create new automations, and it will handle the underlying code and registration for you.
1 maintainer
Metadata
1 maintainer
Release files for openaction 0.1.12
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| openaction-0.1.12.tar.gz | 1.5 MB | Without provenance | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| openaction-0.1.12-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 1.5 MB
Release files / openaction-0.1.12.tar.gz
| Download URL | openaction-0.1.12.tar.gz |
|---|---|
| Size | 1.5 MB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
66b603d260e2e5b8a4f7ecbbc9aca58a7f5ca8d092d61e3bce3e655ae85d793a
|
|
BLAKE2b-256 checksum How to use checksums |
9b593019c609b1884b290c304e2efbde6c7bfc390b128493537b3bc2c54271ef
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Release files / openaction-0.1.12-py3-none-any.whl
| Download URL | openaction-0.1.12-py3-none-any.whl |
|---|---|
| Size | 13.0 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
bbd6b4d2a911312c397ecbf89ad2027c85af620fcf358032a2af4731bb724a22
|
|
BLAKE2b-256 checksum How to use checksums |
e0636a4d809bc9f072bb9e02250d059c8ddea4bfce917fb1b7c2b1b9377bdd1a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Security for openaction 0.1.12
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
No provenance found
None of the files in this release were uploaded with provenance metadata.
A note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
