Process API¶

The Process API is where the AI does the heavy lifting.

This API must be called with a context created by the Create Context API. Each request sends a recent JPEG image of the print. OctoEverywhere's ML image models process the image, then the temporal combination model returns output based on the current image, previous context, and other signals.

HTTP Request¶

Use the ProcessRequestUrl returned by the Create Context API. If that URL fails, switch to the returned FallbackProcessRequestUrl for the rest of the context lifetime.

POST https://<your-process-host>.octoeverywhere.com/api/gadget/v1/process/{ID}

Headers¶

Path Parameters¶

Request Body¶

Send the image as a multipart/form-data POST body. Include exactly one attached file; the uploaded file name does not matter.

This request type is supported by common HTTP libraries in modern languages. The Python SDK shows the expected calling pattern using the Python requests library.

Content-Type: multipart/form-data

Successful Response¶

{
    "NextProcessIntervalSec": 40,
    "PrintQuality": 8,
    "WarningSuggested": false,
    "PauseSuggested": false,
    "Score": 5
}

Response Details¶

NextProcessIntervalSec¶

NextProcessIntervalSec is the minimum time that must elapse before the next Process API call. Your integration can wait longer, but it must not call the Process API sooner than this interval.

Calling after each minimum interval, or as close as practical, gives the temporal combination model the best data for confident print-state decisions.

PrintQuality¶

PrintQuality is the user-facing print quality score from the temporal combination model. It ranges from 1 to 10, where 10 is perfect print quality.

Use PrintQuality for UI status, printer displays, dashboards, or other user-facing print health indicators. Do not use this field alone to trigger warnings or pause actions; use WarningSuggested and PauseSuggested for that.

WarningSuggested¶

WarningSuggested is true when the temporal combination model is confident that there may be a print issue and the user should be informed.

The required confidence can be adjusted with WarningConfidenceLevel when creating the context. This decision uses several signals, so PrintQuality may stay in the 1-3 range for 30-80 seconds before the model has enough confidence to raise this flag.

PauseSuggested¶

PauseSuggested is true when the temporal combination model is confident that the print has probably failed and should be paused.

The required confidence can be adjusted with PauseConfidenceLevel when creating the context. Because pausing a print is intrusive, the model waits for high confidence before raising this flag. PrintQuality may stay in the 1-3 range for 60-120 seconds before PauseSuggested becomes true.

Score¶

Score is the raw temporal combination model score. It ranges from 0 to 100, where 0 is a perfect print and 100 indicates a strong probability of failure.

Use this value for advanced processing, such as smoothing, aggregation, or custom heuristics. Do not use it directly for user-facing status or actions like warnings and pauses.

Error Response¶

If the API does not return a 200 response, it returns an HTTP error code with a common JSON error object.

{
    "ErrorType": "OE_IMAGE_DECODE_FAILED",
    "ErrorDetails": "The uploaded image could not be decoded."
}

Calling Pattern¶

1. Create a context with the Create Context API.
2. Send a JPEG snapshot to the ProcessRequestUrl.
3. Wait at least NextProcessIntervalSec.
4. Send the next snapshot.
5. If the primary process URL fails, switch to FallbackProcessRequestUrl for the rest of the context lifetime.