WhatsApp error 131000: Something went wrong

Code 131000 is the WhatsApp messaging subsystem's own unknown error. Where code 1 is the Graph API saying it could not process the request, 131000 comes from deeper inside the WhatsApp pipeline: the request was accepted, routed to the messaging service, and then failed for a reason the service could not classify. It returns HTTP 500, and no message ID is issued, so the send did not happen.

Meta adds one specific context to the documentation: when you are setting a business public key for WhatsApp Flows encryption, 131000 can mean the signature calculation failed, the internal GraphQL call failed, or GraphQL returned an error. Outside that case, the honest reading is that it is an internal fault.

Because it is unclassified, the first job is to rule out that it is actually a pattern. A single 131000 on a busy number is noise; a stream of them against one recipient, one template, or one media asset points at something on your side the platform failed to describe.

It can appear on both channels: synchronously as an HTTP 500 response body, and occasionally inside a statuses webhook when the failure happens after acceptance. Record which channel each occurrence came from - a webhook-side 131000 means a message ID was issued before the failure, so guard resends with your own idempotency key to avoid duplicates. When a burst of 131000 coincides with elevated code 1, code 2, or 131016 rates, treat it as one platform incident rather than three bugs, and check the status page before touching code; the triage order is summarized in the error-code guide. BSPs cannot see deeper into this error than you can: a provider ticket gets forwarded to Meta with the same fbtrace_id you already hold, so open the Direct Support ticket yourself if you have dashboard access.

Give 131000 a bounded retry with backoff and idempotency, then dead-letter it rather than retrying forever. Alert on clusters of 131000 grouped by template, recipient, and media ID so a systematic cause shows up quickly. Keep fbtrace_id values in your logs; support cannot investigate without them.