Settings and admin
Current runtime assumptions, provider readiness, storage mode, and enabled estimating skills.
Runtime
Skill registry
Provider adapters
The platform is provider-agnostic. Enable whichever LLM adapters you want by setting the matching environment variables, or use the runtime API key box inside a project workspace when you want to test AI actions without rebuilding the app.
Model provider
This is the default runtime path for ingestion, page previews, and baseline document parsing.
Model provider
Recommended for document-grounded reasoning and estimator-facing explanations once configured.
OPENESTIMATOR_OPENAI_API_KEY
Model provider
Positioned as an interchangeable multimodal provider for page reasoning and benchmark workflows.
OPENESTIMATOR_GEMINI_API_KEY
Model provider
Useful for disciplined trade-scope reasoning once credentials are provided.
OPENESTIMATOR_ANTHROPIC_API_KEY
Model provider
This is the modular slot for computer-vision-assisted count, measurement, and markup extraction.
Benchmarks
Benchmark datasets pair plans and reference estimates so we can evaluate classification, CSI mapping, quantity support, and reviewer correction flows.
