Learn how Analyzer in Acrobat consumes generative credits when you upload documents and extract information.
Analyzer uses generative credits to ingest documents and extract attribute values. Consumption depends on the operation and the number of documents processed.
Consumption rates
| Operation | Generative credits consumed |
|---|---|
| Document ingestion | 150 credits per extraction |
| Strict extraction | 50 credits per attribute per extraction |
| Interpretive extraction | 50 credits per attribute per extraction |
| Refresh attribute | Equivalent to the extraction method (Strict or Interpretive) |
How consumption is calculated
Document ingestion and attribute extraction are charged separately. When you upload a document and run an extraction operation, the total consumption includes both charges.
- Consumption for document ingestion is calculated per document file.
- Individual file limits are 150 MB or 400 pages per document.
- Consumption for attribute extraction is calculated per attribute per extraction.
- Refreshing an attribute is a new extraction.
| Example | Calculation | Total consumption |
|---|---|---|
| Upload 10 documents | 10 × 150 | 1,500 credits |
| Run strict extraction on 10 documents with one attribute each | 10 × 50 | 500 credits |
| Run interpretive extraction on 10 documents with one attribute each | 10 × 50 | 500 credits |
| Upload 10 documents, each with one attribute each, and run extraction | (10 x 150) + (10 x 50) | 2,000 credits |
| Refresh one document, with 10 attributes, twice | (10 × 50) × 2 | 1,000 credits |
| Upload one document with 5 attributes, run extraction, and one refresh | 150 + (5 x 50) + (5 x 50) | 650 credits |
Document ingestion
Document ingestion prepares an uploaded file for use in Analyzer. A file can be uploaded either manually or through a linked connector.
Each uploaded file is treated as a separate document for consumption purposes, even when its contents match a previously uploaded file.
Rate: 150 generative credits per uploaded document file.
For example, uploading the same file 10 times consumes 1,500 generative credits.
Strict extraction
Strict extraction identifies information that directly matches the attribute definition.
Use strict extraction when the expected value appears explicitly in the document and does not require interpretation.
Rate: 50 generative credits per attribute per extraction
Interpretive extraction
Interpretive extraction uses generative AI to determine a value from the document’s content.
Use interpretive extraction when the expected value requires contextual analysis or may not appear as an exact text match.
Rate: 50 generative credits per attribute per extraction
Testing and Refreshing attributes
Testing or refreshing an attribute follows the same extraction rules as any other extraction request. Analyzer does not distinguish between extraction operations in this regard.
Testing and refreshing one (strict or interpretive) attribute against one document consumes 50 generative credits.
Consumption rules
- Document ingestion and attribute extraction are charged separately.
- Ingestion consumption is calculated per uploaded (manual, linked, or API) file, not per collection.
- Each upload incurs document-ingestion consumption.
- Uploading a duplicate file does not waive the ingestion charge. Analyzer understands these to be two distinct files.
- Removing and uploading the same file again incurs a new ingestion charge. Analyzer understands these to be two distinct files.
- Generative credit rates don't vary by file type, page count, or file size.
- Attribute extraction (Strict and interpretive) consumes 50 credits per attribute per extraction instance.
- Reprocessing a document consumes additional credits based on the number of attributes being refreshed.
