QA Knowledge Hub

This site replaces PDF-based material with a docs portal that is faster to use in meetings and easier to extend into technical documentation.

Where do JDBIN and JDBON come from?

This should be stated immediately at the start:

1. source data comes from the research corpus and editorial pipeline
2. that is turned into a normalized publishable signal/data model
3. the builder produces two different artifacts from it:
   • JDBIN: a query-optimized binary artifact
   • JDBON: a JSON manifest/control layer
4. an optional public JSON snapshot can also be materialized for the frontend

RSS / corpus / AI-assisted editorial pipeline
-> normalized publishable signals
-> JDBIN builder
   -> JDBIN binary data artifact
   -> JDBON manifest/control JSON
-> optional public JSON snapshot

Important boundary: JDBIN is not just a JSON dump converted into binary. It is a separately built read-optimized segment-based format. JDBON is not the query-heavy payload itself, but the publish, manifest, and version-control layer.

What you can find here

• fast answers to meeting questions
• a separate QA set built directly from jdbin-r2-presentation-2026-07.pdf
• deeper architecture explanations
• benchmarks and limitations
• pitch and pilot material
• a Finnish source version in the default /docs path