Discussions

Hi team, I’m integrating HeyGen into a project workflow and have run into a critical issue that I need clarification on.

When a video job gets stuck at status: "processing" with progress: 0, there is no way to cancel the job via API, and the job may eventually wake up and complete 30–60 minutes later, consuming credits long after my system has already timed out and fallen back.

This creates two major issues:

Unbounded latency, jobs can sit at 0% for 30–60 minutes with no progress signal.

No cancellation mechanism, if I fallback to another provider, HeyGen may still complete later and burn credits unnecessarily.

I tested all logical cancellation/delete patterns:

Endpoint Method Result
/v2/video/{id} DELETE 405
/v2/video/{id}/cancel POST 404
/v1/video_status/cancel POST 404
/v1/video_status.delete?video_id={id} DELETE 404
/v2/video/{id} PUT/PATCH 405/404

So this keeps happening:

Submit a video job (test mode or real mode)

Job enters processing immediately but stays at progress: 0

Job remains stuck for 30-60 minutes

Eventually jumps from 0% - 100% instantly

Credits are consumed at completion

No way to cancel or prevent the late completion

My questions;

Is there an undocumented cancellation endpoint or recommended way to stop a job?

At what exact point are credits deducted, job submission, render start, or completion?

Is there a way to detect whether a job is stuck in a low‑priority queue?

Is there a recommended timeout or best practice for handling long “processing (0%)” states?

Are there plans to expose:

a cancellation API

queue/traffic/priority status

estimated start time

job abort or kill functionality

Thank you in advance for any advice!