Pricing
Three engagement shapes. Every claim in a proposal traces to a measured result, the same rule the platform holds itself to.
Fixed fee
Two-week benchmark of your retrieval stack against our measured baselines: nDCG@10, effective rank, and cost per query on your corpus. You keep the report and the harness.
- Your corpus, our eval harness
- Written findings with reproduction steps
- Go/no-go recommendation
Monthly
A dedicated tenant on the platform: domain-tuned embedding models behind a scoped API, retrained on your cadence with deploy gates enforced before every promotion.
- Isolated workspace (row-level security)
- Deploy gates: no silent regressions
- Usage-metered API with budget ceilings
Annual
The full operating loop (data collection, training, evaluation, serving, and observability) deployed in your cloud, with our team operating it alongside yours.
- Private deployment
- Custom eval suites and SLAs
- Quarterly architecture reviews
Evaluation credits and design-partner seats are sold directly in the store when available. Terms: service terms · privacy.