Useful scope
- Testing explicit inclusion and exclusion criteria on synthetic abstracts.
- Separating absent methods information from a stated exclusion.
- Preparing an uncertainty queue for a human full-text reviewer.
An abstract can justify an exclusion or leave a question open. Apply the written review criteria without filling in methods the abstract never states.
Screening applies the review protocol, not a general impression of whether a paper sounds useful. A relevant topic can still fail an explicit population or publication-type criterion.
Missing information is different from exclusion. An abstract that omits participant age has not established an adult population, but it has not proved the study used children either.
The source review plugin uses TypeSafe questions for title-and-abstract screening. These original software-education examples are a teaching adaptation, not medical advice or a claim that an automated literature review is complete.
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.

Change the evidence, inspect the available labels, then compare the response with the rule.

Use the named fields in the sample. Include the context that distinguishes the outcomes; leave absent facts explicitly unknown.

The decision rule and the choice labels are sent with the evidence. Read the definitions before changing the labels. The model chooses among these supplied alternatives.

The window makes a fresh server-side API request. It displays the highest-probability label, elapsed request time and estimated token cost. A provider failure produces an error, not a missing-evidence label.

Treat the returned label as a proposal. Check the evidence and distribution before consuming it in application code. Download the requests and saved observations. The offline consumer validates their hashes and prints a pending packet without executing it.
| Original input | Observed label | Leading probability |
|---|---|---|
| All abstract criteria | Advance | 96.0% |
| Opinion only | Exclude | 100% |
| Excluded population | Exclude | 100% |
| Age group unknown | Unclear | 100% |
| Protocol absent | Review | 85.0% |
| Contradictory abstract | Review | 98.0% |
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.
No. It means the abstract meets the supplied criteria at this stage. A person still needs to check the full text and the actual protocol.
They should remain available for follow-up under the review process. The label is not a reason to silently discard a record with missing abstract information.
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.