Talk to us

Tell us what your users need to do with video and where your team is blocked. Choose a capability below, or ask us to help you choose.

  • Your project

    Describe the feature or workflow you are building. For pipelines, your answers can draft a starting file.

    Tell us what you need
  • Accounts and plans

    Public signup is off, so Team and Cloud accounts are by request. We arrange Enterprise terms by email.

    Request an account
  • Support

    Help with something you already use, from CLI errors to videos made in ChatGPT.

    Get help
A rough estimate is fine.
Results you needChoose any that apply.
A link to a video like the ones you collect.
Describe the feature, the video you have and the result you need. Tell us where your team is blocked.
Email us

Your starting pipeline file

Built-in steps transcribe the video, sample frames and find scene cuts. Other analysis calls a service you configure through an endpoint step.

Factory's parser checks the file, and it goes with your message as a starting point, not a running pipeline.

Starting pipeline file
{
  "steps": {
    "probe": { "type": "probe" },
    "transcript": { "type": "transcribe", "granularity": { "mode": "segment", "window_ms": 30000 }, "with": { "language": "auto" } },
    "frames": { "type": "sample_frames", "granularity": { "mode": "segment", "window_ms": 10000 }, "with": { "every_ms": 1000 } },
    "scenes": { "type": "detect_scenes", "with": { "threshold": 0.4 } },
    "objects": { "type": "endpoint", "needs": ["frames"], "with": { "endpoint_id": "00000000-0000-0000-0000-000000000000" } },
    "moments": { "type": "endpoint", "needs": ["transcript", "frames"], "with": { "endpoint_id": "00000000-0000-0000-0000-000000000000" } },
    "sentiment": { "type": "endpoint", "needs": ["transcript"], "with": { "endpoint_id": "00000000-0000-0000-0000-000000000000" } },
    "summary": { "type": "endpoint", "needs": ["transcript", "moments"], "with": { "endpoint_id": "00000000-0000-0000-0000-000000000000" } }
  }
}

Factory's parser accepts this file: 8 steps.

objects, moments, sentiment and summary are endpoint steps with a placeholder ID. Before running the pipeline, configure services to do this analysis and replace the IDs. Your service can call the models you choose.

What happens next

We write back to your work email to discuss your project, access and pricing. For Factory, we work through your starting pipeline file with you.

Or email us at hello@editframe.com.