How Long Processing Takes
You've uploaded files and you're trying to figure out when they'll finish processing — whether that "24 hours" you've heard is a hard deadline, whether it applies on weekends, and whether a 500-page statement takes ten times longer than a 50-page one.
Several factors affect how long you'll wait between clicking Start Extraction and seeing a file move to Processed:
Document type and pathway. CSV files — Transaction Lists (CSV) and Check Item Lists (CSV) — are imported synchronously: they are read, validated and imported while you wait, rather than being handed to an external service and queued. There is no external service and no queue for these, so their duration is governed by row count and validation work alone.
PDF files — Bank Statements (PDF), Brokerage Statements (PDF), and Check Images (PDF) — are sent to an extraction service and processed asynchronously. How long that takes depends on the service's current load, the document's page count and complexity, and whether the service is running at all (see below).
Page count and document complexity. A 10-page statement with cleanly formatted tables extracts faster than a 200-page statement with dense layouts, embedded images, or poor scan quality. A statement containing only transaction rows extracts faster than one with graphs, footnotes, multi-account summaries, or irregular page layouts. The relationship between page count and processing time is not linear — a 500-page file does not necessarily take ten times as long as a 50-page file, because layout consistency matters as much as page count.
Current queue depth. The extraction services process files in the order they arrive, with deliberate randomization to prevent starvation. If dozens of large files were uploaded across multiple projects just before yours, your file waits behind them. There is no way to see queue depth from the ALL FILES screen, and there is no priority or expedited processing option available to users or to support.
Time of day and day of week. This is the question customers ask most often, and the answer is that there is no business-hours gate anywhere in the pipeline: nothing pauses, queues differently, or defers work because it is a weekend, a holiday, or the middle of the night. A file submitted on Friday evening is handled the same way as one submitted on Monday morning. What can still vary is how busy the extraction service is at the time, which is not something the platform exposes.
What you can and cannot doYou cannot expedite processing, reprioritize a file in the queue, or pause extraction once it has started. The backend's poll order is deliberately randomized to prevent any one project from starving others, and there is no override available to users or to support.
If a file has been in Processing for days and shows no sign of progress, reload the ALL FILES screen first — the interface does not refresh itself when background work completes, so a stale view can make a finished file look stuck. If it still shows Processing after a reload, contact support with the engagement name, the filename, the document type you selected, and the upload date and time. Support can check whether extraction finished on the backend but the completion signal was never received, and can manually move the file forward if that turns out to be the case.
Support cannot speed up an in-progress extraction, check queue position, or tell you exactly when a file will finish — those are driven by the extraction service's own workload and processing rate, which are not visible to the platform or to support tooling.