Azure AI services
A group of Azure services, SDKs, and APIs designed to make apps more intelligent, engaging, and discoverable.
3,448 questions
Azure Face API v1.2 – Liveness Session Upload Working via /session/start but Status Stuck at "Running"
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(96, 79%, 19%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ENK%3C/text%3E%3C/svg%3E)
Niket Kumar Singh
600
Reputation points
we are currently working on implementing and validating the Azure Face API – Liveness Detection feature in Passive mode to help identify spoofing attempts and verify real user presence and we’re encountering inconsistent behavior while working with the liveness session workflow.
Configuration Details:
- API Version: v1.2
- Region: East US
- Mode:
Passive - Authentication: Valid subscription key and
authTokenreturned during session creation - Upload Method: Using both:
-
PUT /detectLiveness-sessions/{sessionId}/video— intermittent success -
POST /session/start?sessionId=...— consistently returns 200 OK
-
- Session creation (
POST /detectLiveness-sessions) succeeds and returnssessionIdandauthToken. - Video upload using
POST /session/start?sessionId=...returns200 OK.
After successful upload, polling :Content-Type: application/octet-stream Authorization: Bearer <authToken>
returns:GET /detectLiveness-sessions/{sessionId}{ "status": "Running" }- In some sessions, even with correct video upload, status never transitions to
"Succeeded"or"Failed".
"Running"state indefinitely after successful video upload? Are there video format or codec requirements beyond.webm/VP8that might cause backend processing to fail silently? Is there a way to debug or view audit logs for a session? (e.g., via/auditEntriesendpoint?) Are there any backend limitations when using the manual REST API workflow (Postman) instead of the official client SDKs (iOS/Android/Web)? Would switching toPassiveActivemode help with stability in browser/server-side implementations? - In some sessions, even with correct video upload, status never transitions to
- We are not using the frontend SDK for liveness (no Android/iOS/Web SDKs), only backend REST APIs.
- Session auth token is correctly passed within 10 minutes (within validity).
- Tried both
PUT /videoandPOST /session/startfor upload.
We’ve already reviewed and followed the official documentation and REST API reference, but the issue persists.
Looking forward to your guidance and suggestions.
Azure AI services
{count} votes
-
Prashanth Veeragoni 4,440 Reputation points Microsoft External Staff Moderator2025-04-16T16:16:58.3833333+00:00 I Understand that your Session starts and video uploads successfully, but the session stays stuck in "Running" instead of changing to "Succeeded" or "Failed".
This usually happens due to a few common reasons, and here’s how you can fix it clearly.
First, even though your video is in .webm format using VP8, Azure’s backend may silently fail to process it if the resolution, frame rate, or codec settings are not within expected limits. To make sure the video is compatible, you should re-encode it using ffmpeg with standard settings. For example, use this command:
ffmpeg -i input.webm -c:v libvpx -b:v 1M -an -vf scale=640:480 -r 25 output_fixed.webmThis will create a lightweight, clean video file that Azure can process reliably.
Next, even though your token is valid and you’re uploading the video within 10 minutes, small timing mismatches or incorrect headers can cause the video to not get linked properly to the session. Ensure you’re using the header Content-Type: application/octet-stream and that the Authorization header contains the correct bearer token received during session creation. Try using the PUT /detectLiveness-sessions/{sessionId}/video endpoint instead of POST /session/start, as it’s found to be more stable in many cases.
Also, keep in mind that you’re using the REST API directly without Azure’s official SDKs. That’s perfectly fine, but the SDKs (like Android, iOS, or Web SDK) manage session linking and validations internally, which may help avoid backend failures. Just for testing, it’s a good idea to try the official SDK once to see if the same issue happens. If it works via the SDK, the problem is likely with the video formatting or manual upload handling.
Azure doesn’t provide logs or error details directly for the liveness sessions, so if the issue continues even after these steps, the best option is to raise a support ticket in Azure Portal. Provide your session ID, timestamp, region (East US), and a copy of the video used so they can investigate on their side.
Hope this helps, do let me know if you have further queries.
Thank you!
-
Niket Kumar Singh 600 Reputation points
2025-04-17T05:45:07.0633333+00:00 Thanks for the suggestion we tried still 404 error.
-
Prashanth Veeragoni 4,440 Reputation points Microsoft External Staff Moderator
2025-04-18T05:29:47.1233333+00:00 Thanks for the update. If you're getting a 404 error, it usually means that the endpoint you're calling is not recognized — either the URL is incorrect, the session ID is invalid or expired, or the method being used (PUT vs POST) doesn’t match Azure's expectations for that specific endpoint.
Here are a few things to double-check:
· Validate the session ID: Make sure the sessionId used in the URL is exactly the one returned from the POST /detectLiveness-sessions call. It should be active and not expired.
· Endpoint correctness: Ensure the upload call is made to:
PUT https://<region>.api.cognitive.microsoft.com/face/v1.2/detectLiveness-sessions/{sessionId}/videoReplace <region> with your actual region (like eastus) and make sure v1.2 is included in the path.
· Headers and Auth: Double-check that the Authorization header contains a Bearer token, and not your subscription key for this step. Also ensure the Content-Type is set to application/octet-stream.
· Upload method: Make sure you’re using PUT (not POST) for /video upload. Using POST instead will result in a 404 or unexpected behavior.
· Auth token validity: The token returned from session creation is valid only for 10 minutes. If it’s reused after expiry, the backend may not recognize the session.
Hope this helps, do let me know if you still have any further queries.
Thank you!
-
Niket Kumar Singh 600 Reputation points
2025-04-18T05:56:47.4166667+00:00 Thank you for the detailed guidance.
We've carefully revalidated all the steps based on your suggestions, including:
Re-encoding the video using
ffmpegwith the exact recommended settings (640x480,25 fps,no audio, usinglibvpxorlibx264)Confirming the use of
PUTfor/detectLiveness-sessions/{sessionId}/videoas requiredUsing the correct headers (
Authorization: Bearer <token>andContent-Type: application/octet-stream)Ensuring the session ID is correctly referenced and the token is used within its 10-minute validity
Verifying that the endpoint is structured properly with the correct region (
<region>.api.cognitive.microsoft.com) and version (v1.2)To rule out regional support issues, we also provisioned a new Face API resource in the West US region — however, the issue still persists even with this new resource.
Despite following all best practices, we are still encountering the same issue — the session status remains "Running" indefinitely or returns a
404on upload.Alod tried both livenessOperationMode": "Passive" and "Passive-Active"
To expedite the resolution, we’ve also opened a formal support case with Microsoft: Tracking ID: 2504160030005702
Would it be possible to connect on Microsoft Teams for a brief discussion? We believe a quick sync may help spot any overlooked details and potentially unblock us.
Thanks again for your time and support!
-
Niket Kumar Singh 600 Reputation points
2025-04-18T06:04:14.93+00:00 Thank you for the detailed guidance.
We've carefully revalidated all the steps based on your suggestions, including:
Re-encoding the video using
ffmpegwith the exact recommended settings (640x480,25 fps,no audio, usinglibvpxorlibx264)Confirming the use of
PUTfor/detectLiveness-sessions/{sessionId}/videoas requiredUsing the correct headers (
Authorization: Bearer <token>andContent-Type: application/octet-stream)Ensuring the session ID is correctly referenced and the token is used within its 10-minute validity
Verifying that the endpoint is structured properly with the correct region (
<region>.api.cognitive.microsoft.com) and version (v1.2)To rule out regional support issues, we also provisioned a new Face API resource in the West US region — however, the issue still persists even with this new resource.
Despite following all best practices, we are still encountering the same issue — the session status remains "Running" indefinitely or returns a
404on upload.Alod tried both livenessOperationMode": "Passive" and "Passive-Active"
To expedite the resolution, we’ve also opened a formal support case with Microsoft: Tracking ID: 2504160030005702
Would it be possible to connect on Microsoft Teams for a brief discussion? We believe a quick sync may help spot any overlooked details and potentially unblock us.
Thanks again for your time and support!
-
Prashanth Veeragoni 4,440 Reputation points Microsoft External Staff Moderator
2025-04-22T17:36:02.5+00:00 I hope your issue has been resolved with this support request #2504160030005702. And if you have any further queries do let me know.
Thank you!
Sign in to comment
Sign in to answer