Building, integrating, or customizing apps and workflows within Microsoft Teams using developer tools and APIs
SpeakerAttributionNotAllowed on the /content endpoint means Microsoft Graph still considers speaker-attributed transcript content unavailable for that tenant request.
What to verify:
- Confirm Graph transcript access is enabled at the tenant level.
- In Teams admin center: Meetings > Meeting settings > Transcript API access
- Microsoft Graph access must be On.
- If this is off, all transcript requests return
403 ForbiddenwithGraphAccessToTranscriptsDisabled.
- Confirm speaker attribution is enabled.
- In the same settings area, select Configure.
- Include speaker attribution must be On.
- Confirm the request is actually asking for the attributed format.
- The
/contentendpoint supports:-
text/vtt= speaker-attributed -
application/vnd.microsoft.graph.transcript+text= unattributed
-
-
text/vttcan be selected with theAcceptheader or$format. -
application/vnd.microsoft.graph.transcript+textmust be selected with theAcceptheader only.
- The
- Branch on
innerError.code.-
SpeakerAttributionNotAllowedspecifically means the request asked fortext/vttwhile speaker attribution is not allowed for that tenant request. -
GraphAccessToTranscriptsDisabledmeans transcript API access itself is disabled and there is no request-side workaround.
-
- Use the unattributed format as a fallback.
- Retry the same
/contentrequest with:Accept: application/vnd.microsoft.graph.transcript+text - That returns transcript text without
<v Speaker>tags.
- Retry the same
Important limitation:
- The documentation states that the
application/vnd.microsoft.graph.transcript+textformat and the related tenant administrator controls take effect at the end of July 2026. - If the tenant setting appears enabled but
text/vttstill returnsSpeakerAttributionNotAllowed, the supported behavior from the API is still to treat that response as authoritative and retry with the unattributed format. - Metadata requests are not blocked by the speaker-attribution setting; this applies to
/contentonly.
Expected behavior for attributed content:
-
text/vttsucceeds only when tenant speaker attribution is allowed. - If not allowed,
/contentreturns:-
403 Forbidden -
innerError.code = SpeakerAttributionNotAllowed
-
- The documented recovery is to request:
-
Accept: application/vnd.microsoft.graph.transcript+text
-
References: