Policy layer 1
Pairing
Controls whether an unknown sender is approved for a DM path.
Allowlists
Control who or what places are allowed.
OpenClaw channels
Channel setup guide
Use this page when the goal is to get OpenClaw into a real chat channel without collapsing credentials, policy, and reply behavior into one vague setup step.
Core idea
A useful setup is more than “the bot is online.”
Shared setup shape
Most channel setup follows one repeated pattern
- connect the channel with working credentials
- make sure the Gateway runtime can actually read those credentials
- decide direct-message policy
- decide shared-space policy
- decide whether ambient shared-space replies require a mention
- test DM and shared-space paths separately
Three policy layers
The controls people most often mix up
Policy layer 2
Policy layer 3
Mention gating
Controls whether the bot replies ambiently in shared spaces.
Safe starter defaults
Start narrow, then widen on purpose
- runtime-readable credentials
- pairing for DMs
- allowlists for shared spaces
- mention gating on
- broaden access only after both DM and shared-space tests succeed
Next route
Useful next page
Shared pattern page
Cross-channel setup patterns
Learn the reusable mental model before you split into platform-specific docs.