Lender knowledge
Lender knowledge, structured for retrieval.
Criteria answers, document requirements and reference data, chunked, labelled and versioned. Built for grounding copilots and agents that must answer from source rather than from memory.
What you get
- Criteria answers labelled by lender, product type and topic
- Document requirements and packaging rules as structured records
- Chunked text built for retrieval pipelines and grounding
- A version and source reference on every answer
- Reference data for lenders, product types and property categories
GET /v1/lenders
GET /v1/lenders/{lenderId}/criteria
POST /v1/criteria/search Endpoints gated. Full schemas in /openapi.yaml
Use cases
Where it earns its keep.
Grounded copilots
Give an assistant the lender's own words with a version and source attached, so its answers survive a compliance review.
Criteria pre-checks
Search criteria across lenders before a decision in principle, and show the basis for every match.
Packaging without the binder
Pull document requirements per lender and case type into your workflow instead of a shared drive of PDFs.
Data and audit
Answers are extracted from lender-published material, labelled and versioned. Where a source changes, the prior version is retained and the change is visible.