Capabilities

Last verified: 2026-07-26 against dev at f0388abb.

This is the product capability map. It is higher level than the API and component READMEs.

Assessment and Document Workflows

• Create and list assessments.
• Upload documents to an assessment.
• Validate document uploads.
• Track document processing status and completeness.
• Remove documents or assessments.
• Rerun extraction for a document.

Extraction and Review

• Register schemas and extractors.
• Generate processors from schema/extractor combinations.
• Create batches and batch runs.
• Run extraction with document-specific processors.
• Review extracted items and fields.
• Validate extracted items individually or in bulk.
• Record corrections.
• Export extracted data through named exporters.

Extractor Platform

• Azure Function extractor apps.
• Service Bus queue based routing by extractor name.
• Local Docker Compose support for backend, frontend, database, blob storage, Service Bus, and function apps.
• Dummy, Docling, Azure Document Intelligence, Excel, and DOCX extractor paths represented in local/development docs.
• Smart chunking support for large ADI documents.

Tenancy and Security

• Organization-scoped assessments and derived records.
• Route-level and storage-level organization access guardrails.
• Frontend organization switching safeguards.
• BYOK LLM key design for consumer-owned LiteLLM virtual keys.

Evaluation

• Phoenix/Arize integration for trace, dataset, experiment, and evaluator workflows.
• Extraction accuracy evaluators.
• Checklist/report accuracy experiments.
• Smart chunking smoke-test and benchmark tooling.

Infrastructure and Operations

• Terraform-managed infrastructure outside this repo.
• Docker Compose local development.
• Backend, frontend, extractor, Phoenix, and notebook-specific setup docs.
• Ad hoc migration/backfill script docs for known operational tasks.