Invoke Automation Statelessly
How to invoke an Automation without inputs and outputs being persisted by Guru
Invoke Automation Statelessly
It is possible to invoke an automation statelessly, such that no data is persisted server-side. This ensures that neither the input, nor the output, are kept by Guru. This endpoint imposes some extra restrictions in order to remain stateless:
- Only supports a single file as input. No text input can be provided.
- File is limited to a maximum size of 20Mb.
- The call must be performed synchronously, meaning that the API will not return until all processing has completed.
- A maximum of 100 concurrent requests may be in-flight at any given time. Attempting to start more may result in a
429 Too Many Requests
error response. - If lost, the output cannot be re-fetched again later.
Request
The FormData must contain a single property called file
, which is the byte-content of the file.
Response
The output of each document in the request. This will contain more than 1 object only if Guru detected sub-documents within the request document.
The unique identifier for this invocation.
The output of the invocation, encoded as a string. If the output of the automation is configured to be JSON, then this String will be parseable as JSON.
The reason for failure, if the invocation failed.
The status of the invocation. Either “Success” or “Failed”.