Run a local gateway
Install and start the OpenClaw gateway on your desktop or a small home machine such as a Mac mini or Raspberry Pi. Your conversation history and active skills stay on your hardware.

Self-hosted open source assistant powered by messaging integrations.
By OpenClaw Foundation

Install and start the OpenClaw gateway on your desktop or a small home machine such as a Mac mini or Raspberry Pi. Your conversation history and active skills stay on your hardware.
Link messaging apps like WhatsApp, Telegram, or Discord to send text prompts, photos, or voice notes directly to your agent. Replies arrive in the same chat without opening a separate dashboard.
Set up timed reminders and recurring prompts so the gateway checks in or runs routine tasks on a schedule you define.
Link your existing AI model subscriptions or endpoints and explicitly authorize which local tools or community skills the agent can invoke.
MIT-licensed software. Hosting and external model charges may apply separately.
OpenClaw is MIT-licensed software with no OpenClaw subscription. Running the gateway may incur hosting or model-provider charges, depending on the setup.
The OpenClaw gateway runs on your own computer or server, keeping conversation history and active skills on your hardware. However, if you connect a cloud model provider rather than a local model, that provider still processes the requests you send.
Link a supported messaging app such as WhatsApp, Telegram, or Discord to your OpenClaw gateway. You can send text prompts, photos, or voice notes directly in chat, and the assistant replies in the same conversation thread.
Yes. You can configure timed reminders and recurring prompts in the gateway so OpenClaw checks in or runs routine tasks on your schedule, such as triggering an authorized test-runner skill and posting results to Discord.
Yes. OpenClaw supports policies that allow or deny tools, as well as sandbox settings for supported operations. Check both when deciding whether an assistant should be able to read a workspace or run local commands.
You can install and start the OpenClaw gateway on Mac, Windows, or Linux hardware, including a desktop, Mac mini, or Raspberry Pi. You then connect your existing model subscriptions, hosted endpoints, or local models to power responses.
The schedule runs in the OpenClaw gateway. Keep the machine hosting that gateway available when the job is due. If the job should send a message, configure its delivery channel as part of the schedule.
Public experiences and discussions, with links to the originals.
No discussions collected yet.
Explore 7 products that share at least one second-level category.
View all 7 alternativesFound something incorrect? Suggest a correction for our editors. It’s free; no account or email is needed.