Proof · single tender
One endpoint, one alert. Paste a normalised tender JSON on the left and a Company Profile on the right — score, verdict and the 3-line rationale come back and persist to the audit table in under a second. This is the Bidloom scoring engine end to end.
Input · 2 JSON blobs
Both blobs are validated server-side against a shared contract. A wrong shape 400s with field-level errors — no silent drift, no mystery masks.
/api/matchscoreAudit log · last 5
Loading audit log…
Try a different notice
Three sample tenders to drop into the Tender box — copy any of them and watch the score swing on industry, value and deadline.
IT endpoint refresh — phase 2
Yorkshire · £150,000–£180,000
Building fabric maintenance — south zone
East Midlands · £380,000–£460,000
Security-cleared FM helpdesk (lot 4)
Multi-region · £2,000,000–£2,800,000
Every score is auditable: the inputs you typed and the 5-line sub-score breakdown are persisted in TenderMatch the instant a result returns. The weight model is fully deterministic — change the Profile, regenerate, and the same Tender scores reproducibly.
Five sub-scores, five weights.
Every Tender is decomposed into five sub-scores — Industry & certs, Contract value, Geography, Deadline pressure, and Competition. Each is weighted and combined into a 0–100 score with a deterministic bid/review/skip verdict. Every row pulled from the API is reproducible from the same inputs — no model drift.