Back to blog

How to run OpenCode from Slack with Pipa

Connect OpenCode on your computer to Slack with the free, open-source Pipa CLI, then send your first request from a trusted channel.

Posted by

I use OpenCode for work across Lunch Pail Labs, but I do not always want to sit in front of the terminal to reach it. Pipa connects the OpenCode running on my computer to Slack, so I can give it work and continue the conversation from a channel.

Pipa Local is a free, open-source bridge. It uses your existing OpenCode setup and works inside the folder you choose. This tutorial walks through the complete setup in under four minutes.

Prefer YouTube? Watch the OpenCode Slack tutorial here.

The short version

npm install --global @usepipa/pipa
pipa init
pipa start

Run pipa init inside the folder you want OpenCode to access, follow the Slack setup wizard, then keep pipa start running.

What you need

  • Node.js 22 or newer.
  • OpenCode v1 installed and authenticated.
  • A Slack workspace where you can create and install an app.
  • A folder containing only the files OpenCode should be able to reach.

You can check the current requirements on the Pipa Local install page.

1. Install the Pipa CLI

Install Pipa globally from your terminal:

npm install --global @usepipa/pipa

This gives you the pipa command used for setup and for starting the Slack bridge.

2. Choose the folder OpenCode can access

Move into the folder where you want the Slack agent to work, then start the setup wizard:

cd /path/to/your/folder
pipa init

I use a container folder called studio for the projects I want my agent to reach. You can use one project folder instead if you want a smaller boundary.

Choose the folder and Slack channel carefully

Anyone in a connected Slack channel can ask OpenCode to read or change files in this folder. Use a trusted channel, invite only people who should have that access, and do not initialize Pipa higher in your file system than necessary.

3. Follow the Slack setup wizard

The wizard asks you to name your bot and confirm its working folder. It then opens Slack so you can choose a workspace and create the app from a generated manifest.

Slack requires two tokens:

  1. An app-level token with the connections:write scope.
  2. A bot user OAuth token created after you install the app in your workspace.

Paste each token into the wizard when it asks. Keep both tokens secret. Pipa saves the configuration on your computer.

4. Start the OpenCode Slack bridge

Once setup is complete, start Pipa:

pipa start

Keep this process and your computer running while you use OpenCode from Slack. Pipa Local is not a hosted service, so the connection stops when the process or computer stops.

5. Invite the bot and test it

Open a trusted Slack channel, invite the bot, and mention it with a small request. I start with a simple message to confirm that the connection works before asking it to read or change anything.

@your-agent Hey there, how are you doing?

The OpenCode configuration and skills available in that local instance are now available through Slack, within the folder boundary you selected.

What Pipa Local does

  • Connects Slack to the OpenCode installation on your computer.
  • Uses the OpenCode configuration and skills you already have.
  • Gives you control over the folder, channels, tokens, compute, and uptime.

It does not turn your computer into a hosted, always-on autonomous agent. If you stop Pipa or close the computer, it goes offline.

Run OpenCode from Slack

If you already use OpenCode and want to reach it from Slack, start with one trusted channel and one carefully chosen folder. You can install Pipa Local for free or inspect the source in the Pipa GitHub repository.

How to run OpenCode from Slack | Pipa