All trained models in the registry. Champion is the one currently serving new scoring requests. Click Report to open the per-model training report (AUC curves, decile lift, feature importance).

Engineered features the scoring pipeline computes at run time when a vendor file doesn’t supply them directly. Three flavors: expression (pandas formula), aggregate (groupby from a DB table), lookup (key-based join against a reference table). Inputs go through the alias dictionary, so an input named customer_id resolves from any vendor column that maps to it.

Draws from your scoring runs become curated mail lists. Build one, optionally route through approval, download the CSV for your mail vendor, then mark it exported.

Per-state suppression status across your planning horizon. Red = inside an election suppression window. Sand = 6–10 weeks pre-election (peak power). Teal = 1–3 weeks post-election recovery. Grey = clean. Dots mark elections; vertical bars mark the federal general. Pick a target drop date to overlay it.

Uploaded addresses are normalized (USPS-style) and matched against marketing list pulls by (address_norm, zip5). Duplicates are skipped automatically.

KMeans on standardized features. Claude names each cluster from its centroid and z-deviations. Works best on a labeled file (with a RESPONDERS column) so cohorts get a real response rate.

Tokens for scripts, pipelines, and external tools to push and pull from the platform. Use as Authorization: Bearer sndp_… in your requests. Permissions inherit from the creating user’s role. Plaintext is shown only once at creation — store it somewhere safe.