Actions21
- Azure OpenAI Actions
- Batch API Actions
- Content Safety Actions
- Document Intelligence Actions
- Language Actions
- Speech Actions
- Translator Actions
- Vision Actions
Batch API → Run Full Batch
AI-generatedSummary
Run a complete Azure OpenAI batch inference job by uploading a JSONL file, starting the batch, polling its status until completion, and downloading the results automatically.
Inputs
- Authentication (required) — Choose between API Key or OAuth2 authentication for Azure.
- Binary Property (required) — Name of the binary property containing the JSONL file to upload.
- Batch Endpoint (required) — Select the Azure OpenAI endpoint path for the batch job, e.g., Chat Completions, Completions, Embeddings, or a custom path.
- Custom Endpoint Path — If 'Custom' is selected for Batch Endpoint, specify the custom relative endpoint path here.
- Completion Window (required) — Maximum time window for the batch to complete. Currently only '24 Hours' is supported.
- Polling Interval (Seconds) (required) — Seconds between each status check while polling for the batch job completion.
- Output as Binary (required) — Choose whether to return the raw JSONL output file as binary or parse each line into JSON objects.
- Binary Property Name — If output as binary is selected, name of the binary property to store the downloaded JSONL file into.
Output shape
a single item per input containing either parsed JSON result objects or the raw JSONL file as binary data.
If 'Output as Binary' is true, the output includes a binary data property containing the JSONL file; otherwise, the output JSON contains a 'results' array with parsed JSON objects from each line, plus metadata such as batchId, inputFileId, and outputFileId. The operation throws an error if the batch fails or does not complete within the specified completion window.