Standalone automation
Complete provides queues, workers and a workflow ledger. Control operates configured Complete/Node hosts through Remote Ops.
Follow work in the ledger →USE CASES
Useful data starts
with a specific question.
Which prices changed? What does this saved page contain? Why did yesterday's collection stop? Here is how Guida fits into the work.
These are illustrative workflows, not customer endorsements or ready-made extractors. Each source needs its own implementation and checks.
Track a changing product catalog
You need the same fields from a set of product pages: an identifier, name, price, currency and availability. Copying them manually is slow, and small layout changes can break an unattended script.
-
Inspect one page
Use Complete to inspect the rendered page, or configure an HTTP/browser source in Enterprise.
-
Extract a sample
Write the source logic and check the actual records, including missing values and types.
-
Define the output
In Studio, choose a stable record key, map fields and preview the API response.
-
Prepare repeat runs
With your operator, release the source and configure authorized execution. Inspect failures when the source changes.
Useful outcome: a maintained extraction process with a defined record shape. Historical comparisons require an explicit storage and comparison design; a current-record API alone is not a price-history service.
Extract from material you have already collected
You have saved documents and want to change the fields you extract without visiting the original website again. Keeping collection and processing separate makes that iteration easier to inspect.
Enterprise includes a Linux document worker for retained-document processing. Configure its extractor for the fields you need, then pass the resulting records into your review workflow. It runs separately from the Windows browser hosts, so processing can use saved material without reopening the original page.
Useful outcome: reusable extraction logic over retained inputs, with records that can feed a reviewed data workflow. Document-format support and the chosen extractor must be established for the evaluation; this is not a promise to parse every file automatically.
Find out why a collection stopped
A recurring collection should leave enough information to distinguish completed work, a failed source and work that has not run yet. Operators need to inspect the failure before choosing a recovery action.
Enterprise execution
The Console exposes campaigns and fleet operations for enrolled hosts. Work follows the configured authorization, capabilities and reviewed release.
Read the operator guide →Useful outcome: a clearer account of what ran and what needs attention. Source changes, access restrictions and transient failures still need a diagnosis.
Learn the loop before connecting a real source.
Try the synthetic careers-feed exercise in your browser. Inspect a relevant role and its evidence, change the input, then repair the extraction. Download the same code to continue locally or use it in a prepared Studio workspace.