Control systemd timers/services (enable/disable, start/stop, and timer cadence). We keep these quiet unless you are running an active multi-agent session.
Controls update notification behavior and shows latest available release info.
Mode
Max level
Status
(load)
Latest release (raw)
Together.ai (OpenAI-compatible)
Ops-only config to validate calls. This stores a Together API key + default model so we can later route requests through it (e.g. Qwen2.5-Coder).
Base URL
Serverless uses https://api.together.xyz. Don’t paste a /models/… dedicated-endpoint URL.
Model
Or type a model override:
API key
Saved to: /var/lib/clawdzo/console-data/together-config.json
Test
Prompt
B2B DM (Bridge)
Cross-box notes + newborn instructions. Pair new boxes to the boss here.
Boss token
On the boss box, click Copy. On a new box, paste the token into Peer token and Save.
(Masked in UI; copy grabs full value.)
Peer
No sudo needed; stored file-backed in DATA_DIR.
Peer URL
Peer token
Live: (starting…)
Direction
Summary
Text
Recent
Browser
This tab exists because we hit an intermittent failure where a stale Chromium process was already listening on the CDP port and the managed browser failed to start.
The long-term fix is: make the CDP port explicit per box, and keep the "what to do" steps written down here.
Current settings
Profile
CDP port (Chromium remote debugging)
Writes to: /root/.clawdbot/clawdbot.json (browser.profiles.<profile>.cdpPort)
What happened + what to do (copied from bot_19cd0441f44666b2d)
That error is exactly what it says on the tin: something was already listening on 127.0.0.1:18800, so when Clawdia tried to start a new Chromium CDP session for profile clawd, it failed and the browser control endpoint returned HTTP 500.
Fast recovery:
1) Check what’s holding the port:
sudo ss -ltnp | grep :18800 || true
2) Kill whatever is bound to it (usually Chromium):
sudo fuser -k 18800/tcp || true
# or:
sudo pkill -f 'remote-debugging-port=18800' || true
3) Restart the gateway so it re-spawns the managed browser cleanly:
sudo clawdbot gateway restart
Long-term: set a non-default CDP port per box, and/or implement "probe then attach; else kill+retry" on EADDRINUSE.
Clawd
UI branding only. This does not rename services/hosts; it just changes how the Console refers to the assistant.
Assistant name
Saved to: /var/lib/clawdzo/console-data/clawdreadme.json
Current
Loading…
Tip: after changing the name, refresh open tabs to see it everywhere.