Troubleshooting
Check availability, permission, agent integration and transport.
Sessions are visible but sending fails
Check that the host provider is ready. If Codex App was launched independently, ReignHub may only be reading history. Wait for the task to finish, quit the App and launch through ReignHub so both connect to the same app-server.
Codex CLI sessions owned by another backend can also have writer conflicts. Shared history files do not establish control of an active session. See Codex CLI.
Device appears offline
Check sleep settings, whether ReignHub is running, network access to the service and whether both sides use the same service root URL. Enterprise Wi-Fi may isolate peers even on the same SSID.
Do not delete local data to refresh the network: this loses pairing identity. Restore connectivity and reconnect; the client catches up to current message state.
Direct connection fails with no fallback
Check relay entitlements. Official Free permits direct connections only. Self-hosted services can use their own relay; verify Worker WebSocket routes and Durable Object bindings.
Browser copying, pairing or encryption fails
Use HTTPS, or localhost for local development. Check clipboard and site-storage permissions. Private browsing, storage quotas and extensions may also affect persistence.
Self-hosted probe has no effect
Open both 26330/TCP and 26330/UDP, and use DNS-only records. Ordinary Nginx HTTP proxying does not support these probe protocols. Check the Worker SELF_STUN and TCP_STUN settings.
Report an issue
Include client, OS and agent versions, service type, direct/relay status, reproduction steps and redacted errors. Do not upload passwords, pairing keys, cookies, OAuth tokens or complete private conversations. See Contributing.