Skip to main content
Every JobInput accepts an optional config block. If you omit it, both SDKs inject defaults that match the values in models.py and sdk/typescript/src/models.ts. Override only what you need to keep jobs predictable and repeatable.
config.documents and config.video map 1:1 across the Python and TypeScript SDKs. You can mix overrides from both sections in the same job without providing the entire structure.
1

Decide what needs tuning

List the behaviors you want to change (for example, reprocessing previously seen PDFs or sampling video frames more frequently). This keeps your overrides minimal.
2

Override the relevant section

Set config.documents or config.video fields directly inside the job payload. Defaults fill in any fields you skip.
3

Submit and verify

Send the job through the SDK or REST API. Monitor status using the Jobs endpoints to confirm the config produced the expected behavior.

Document config

Control reprocessing and element extraction for PDFs, office docs, and text.

Video config

Adjust screenshot cadence, sensitivity, token budgets, and advanced SSIM thresholds.
import type { JobInput } from "@trelent/data-ingestion";

const job: JobInput = {
  connector: {
    type: "url",
    urls: ["https://signed.example.com/sample.pdf"],
  },
  output: { type: "s3-signed-url" },
  config: {
    documents: {
      reprocess_documents: false,
      extract_elements: true,
    },
    video: {
      screenshot_interval_seconds: 2,
      sensitivity: 0.2,
    },
  },
};