Troubleshooting

The problems that actually come up, and where to look first.

SIP trunk will not register

Check, in this order:

  1. Credentials. Retype rather than paste — a trailing space in a SIP password

is the single most common cause.

  1. IP allowlist. If your carrier authenticates by IP, our media server

addresses must be on their allowlist. Ask us for the current list rather than inferring it; a deployment can span several media servers.

  1. Host. Some carriers give a different host for registration than for

termination. Use the one their SIP connection page shows.

If registration is up but calls still fail, the problem is routing rather than authentication — go to the next section.

Outbound calls fail immediately

  • Caller ID. Most carriers reject a call presenting a number you are not

authorised to use. Check the caller ID on the campaign.

  • Number format. Use E.164 — +15551234567. Some carriers accept other

formats; relying on that is fragile.

  • Trunk selection. If the deployment has more than one media server, a

campaign can select a trunk that is not provisioned on the server handling the call. The platform falls back, but the result is calls leaving through the wrong route. If outbound routing looks wrong, check with us which server is handling your calls.

Outbound calls give up too quickly

Ring timeout is configured per deployment. The platform sets it explicitly rather than inheriting Asterisk's 30-second default, but if calls are being abandoned before a reasonable person would pick up, this is the setting to raise.

Inbound calls do not arrive

  1. Is the number added? Admin → Telephony → Phone Numbers.
  2. Is inbound enabled on it? A number can exist for outbound caller ID only.
  3. Is there a route? A number with no routing profile has nowhere to send the

call.

  1. Check the routing log. Admin → Telephony → Routing Log records every

decision. If the call reached us at all, it is in there.

If the routing log has no entry, the call did not reach the platform — that is a carrier-side routing question.

The agent is stuck on a call that already ended

Press End call. Agent presence follows call state, so ending the call releases the agent.

If the agent still shows as on a call, a background reconciler closes stale call records on a schedule and presence will correct itself. If it happens repeatedly rather than occasionally, tell us — that is a defect, not a workflow.

Softphone will not connect

  • Browser permissions. The browser must be allowed microphone access. Check

the padlock in the address bar.

  • Firewall. SIP signalling is proxied over port 443 precisely so corporate

firewalls do not block it. If it still fails, media (RTP) may be blocked — ask your network team about UDP to the media server.

  • Audio device. Check the device selector in the softphone; a disconnected

headset selected as input produces a call with silence on it.

Recording is missing

  • Is recording enabled for the tenant? Admin → Telephony → Recording.
  • Has retention already expired it? Retention is capped by plan — 30 days on

Starter, 90 on Growth, up to 365 on Enterprise.

  • Has your storage allowance filled? With the default overflow policy the oldest

recordings are deleted first.

Transcription or AI features unavailable

AI features require a plan that includes them. Admin → Administration → Subscription shows what your plan covers, and Admin → Configuration → AI Settings shows the configured providers.

If AI is included and configured but calls are not being transcribed, check whether your AI minutes allowance for the period is exhausted.

An AI agent said it booked something that was not booked

It should not. Where the calendar integration is unavailable, the agent is supposed to report the failure rather than claim success, and the attempt should land in Pending Bookings for a human.

If you see the AI confirming a booking that did not happen, report it — that is a defect worth our attention immediately.

A webhook stopped arriving

Check the delivery log first: Admin → Configuration → Outbound Webhooks → Logs. It records what was sent and what your endpoint returned, and lets you replay a delivery. That will tell you within a minute whether the problem is ours or yours.

Still stuck

Contact us with Technical support selected. Include the call ID if it is about a specific call — it is on the call record — and roughly when it happened.

Something here wrong or missing? Tell us — these pages describe the platform as it actually behaves, so a mismatch is a bug we want to know about.