PDF text extraction review: six recorded examples
The downloads contain six original inputs, their expected labels and one recorded API observation per input revision. Disagreements stay visible.
Provide extracted page text and parser observations. Distinguish usable content from an OCR candidate or missing evidence; empty text alone does not establish that a page is scanned.
This is a text-based routing exercise. Paste extracted text and parser facts; the page does not inspect a PDF image.
OCR candidate is a reason to consider an OCR attempt, not a promise that OCR will recover the content. A real pipeline must preserve page order and validate the result.
The linked document router separates readable pages from pages needing OCR. Its third-party measurements are not results for these synthetic samples.
This is an independently authored text demonstration inspired by the linked X discussion. Its rules and synthetic inputs are ours. It does not reproduce the author’s application, verify reported outcomes, or perform an external action.

Use the sample to understand the rule, then test your own input for pdf text extraction review.

Provide extracted page text and parser observations.

Open the decision rules and keep the intended distinction between Usable text, OCR candidate, Review. Edit the choices only when testing your own rule.

Use a sample first, then try your own input. Compare the selected label with the other returned probabilities.

Review Review before using a result. The label is a decision about the supplied information, not an executed action.
| Original input | Observed label | Leading probability |
|---|---|---|
| Readable page | Usable text | 100% |
| Image-only content | OCR candidate | 93.0% |
| Watermark only | OCR candidate | 99.0% |
| Unknown empty extraction | Review | 99.0% |
| Broken character mapping | OCR candidate | 99.0% |
| Conflicting coverage | Review | 94.0% |
These are six original synthetic inputs, tested once per request revision. The percentage is the leading option probability, not measured accuracy. Authored expectations and any disagreements remain in the downloadable evidence. 6 of 6 results matched the authored fixture expectations.Verified
The downloads contain six original inputs, their expected labels and one recorded API observation per input revision. Disagreements stay visible.
The single-request SDK runner uses your private TYPESAFE_API_KEY. The offline consumer reads saved observations, checks request hashes and prints a pending output. It does not call the model or change any external system.
# Offline: inspect the downloaded saved observations. No API key is needed.
python preview_choice_workflow.py samples.json saved-api-results.json
# Optional live request: install the SDK and set TYPESAFE_API_KEY privately.
npm install @typesafe-ai/sdk@0.6.0
node run-choice-sample.mjs samples.json case-1
# The live response prints to the terminal; it does not update saved-api-results.json.The input contract and these limits define what the result means.
Yes. Its body may be a scanned image while only a watermark or header is extractable. The supplied parser facts must describe that gap.
No. The editable example accepts text and parser observations. A separate parser would inspect the file and an OCR service would process selected pages.
It describes the distribution over the supplied choices for this request. These six synthetic examples are functional tests with authored expectations, not a production accuracy benchmark.
No. The demo displays a label and distribution. The downloadable consumer produces a pending preview packet. Application code and a separate authorization policy would control any real action.
Edit the example above to see how a changed fact affects the decision, or open the Playground for another task.