Prompt size at turn 12
45× smallerCrystal footprint vs raw-accumulating at the final turn. Crystal: 1,360 tokens. Raw accumulating: 61,146 tokens — 45× larger and growing every call.
Buy a construction and architecture library built around shared code, specification, permitting, and handoff context before adding project documents.
Three arms, same model (gpt-4.1-mini), same question set — no context, raw-accumulating, hosted crystal. At turn 12 the crystal sent 1,360 tokens; the raw arm sent 61,146 — 45× more, compounding every call.
Crystal footprint vs raw-accumulating at the final turn. Crystal: 1,360 tokens. Raw accumulating: 61,146 tokens — 45× larger and growing every call.
A formatting rule planted at turn 1 was checked deterministically every subsequent turn. No-context retained it through all checkable turns. Crystal: 1 / 2 checkable turns held.
All three arms produced substantive answers through turn 12 with no meaningful quality decline. The separating factors are prompt size and instruction retention, not raw answer quality.
How this was run. 12 turns, Construction and Architecture vertical, gpt-4.1-mini held constant across all three arms. Raw-accumulating arm: source corpus injected at turn 1, then full conversation history re-sent on every subsequent turn — the window grows every call.
Workspace subscription for project teams that need a shared code, spec, and permitting reference layer before project-specific material is connected.
5,500,000 credits/mo
Up to 18M raw-token-equivalent shared built-environment reference coverage.
Raw RAG can inflate every prompt with retrieved text. The crystal keeps shared operating judgment compact and reusable, so teams buy a monthly memory layer instead of repeatedly paying for larger, noisier context windows.
Each plan includes protected access to prepared industry memory across the recurring work areas below. Your team sees the benefit in chat while the library's source construction remains private.
A code-orientation layer for building terminology, permitting questions, and jurisdiction-facing document expectations.
A specification reference layer for material language, scope boundaries, and trade coordination terms.
A project coordination layer for drawing review, RFI language, and construction administration context.