case-study
InQuest Project Binder RAG QA
Project-aware RAG QA matrix for parcel context, site-plan data, binders, generated files, and media attachments.
Overview
InQuest Project Binder RAG QA captures the work behind InQuest/INQI Quest operation logic. A 2026-06-04 source review covered an operation-logic matrix for project metadata, site-plan data, elevation data, Project Binder, Reference Binder, All Projects Binder, vector-store retrieval, manual attachment priority, generated-file saving, PDF export, media attachments, and web-search toggles. The public case study turns that into a product-engineering surface for QA design, context routing, binder persistence, and public-safe evaluation without exposing project files, account data, private locations, or raw generated outputs.
What It Covers
- Defines context precedence across project metadata, site-plan objects, elevation data, and manual attachments
- Tests Project Binder, Reference Binder, and All Projects Binder retrieval modes
- Covers save-to-binder behavior for PDFs, generated files, and media attachment outputs
- Separates public QA architecture from private client files, locations, and generated artifacts
Stack And Topics
- RAG
- Vector Stores
- OpenAI APIs
- Project Context
- Binder Workflows
- PDF Generation
- S3
- QA Matrix
- Web Search
Public Signals
- QA matrix: 20+ scenarios InQuest operation-logic spreadsheet checked 2026-06-04
- Context modes: 3 binders Project Binder, Reference Binder, and All Projects Binder toggles
- Site context: parcel + elevation project metadata, raw site-plan objects, and elevation-data checks
- Public posture: sanitized no private project files, account data, locations, or raw outputs published