Enter the name of your project.

Project Name:

Enter the author of the project.

Project Author:

Specify the version of the project configuration.

Version:

Specify the directory where the input text files are located for review.

Input Directory:

Specify the file name and path where the results will be saved. Ensure the path exists.

Results File Name:

Select the format for the output results. Available options are CSV and JSON.

Output Format: CSV JSON

Set the log level for the process. Low logs minimal information, medium logs more and prints out on screen, and high logs all details also on file.

Log Level: Low Medium High

Choose whether to enable duplication for debugging purposes. This will run model queries twice.

Duplication: No Yes

Enable Chain of Thought (CoT) justification to request and save model reasoning for answers provided.

CoT Justification: No Yes

Choose whether to generate and save summaries of the manuscript.

Summary: No Yes

Describe the persona or role the model should adopt when performing the task.

Persona:

Specify the task the model should perform.

Task:

Outline the expected result in detail, in particular stating you are looking for a JSON response object.

Expected Result:

Provide any specific definitions to avoid misunderstandings.

Definitions:

Include examples to clarify the request.

Examples:

Specify a failsafe response if the concepts are unclear or missing.

Failsafe:

Optionally document this review as a data-extraction stage in a shared RevAIse review record. Disabled by default; normal review outputs are unchanged when enabled.

Enable RevAIse: No Yes

Path to the RevAIse record to create or update. Reuse the same file across stages to build one cumulative record.

Record File:

Record format. If omitted, it is detected from the record file (JSON by default).

Format: JSON YAML

RevAIse schema version to record.

Schema Version:

How much a human reviews the AI output. Defaults to None, since prismAId itself performs no human review; raise it to reflect the review you actually carry out.

Human Oversight Level: None Minimal Exception only Confidence based Sample review Full review

Human-readable label for this extraction stage.

Stage Label:

Run identifier. Reuse the same id to update the same run; change it to record a new one (e.g. a pilot vs. a full extraction).

Run ID:

Human-readable label for this extraction run.

Run Label:

Extraction form identifier, name, and version.

Form ID:
Form Name:
Form Version:

Identifier of the extractor performing this run.

Extractor ID: