Database Reference

The Brand Match App has no local database. All data lives in a shared Neon Postgres instance, accessed via the ca-data API backend. This page documents the key Neon tables that power the app.

---

Entity Relationship Diagram

---

matches

The central table powering all match views. One row per creator–brand pair.

Field	Type	Notes
id	Int	Primary key
brand	String	Brand name
creator	String	Creator name
agent	String	Assigned agent name (annie, charlie, henry, or null for unassigned)
match_score	Float	Overall match score 0–100
rate	Int	Creator's rate in dollars
cpm	Float	Creator's CPM
avg_views	Int	Creator's average view count
subscribers	Int	Subscriber count
niches	String	Comma-separated niche tags
platform	String	Primary content platform
status	String	Pipeline status: not_pitched, reached_out, in_negotiation, won, not_a_fit
pitch_narrative	String?	Pre-written pitch angle for this pair
pitch_angle	String?	Shorter pitch hook
notes	String?	Agent notes
us_pct	Float	US audience percentage
male_pct	Float	Male audience percentage
brand_deals_total	Int	Number of past brand deals for this creator
niche_fit	Float	Signal 2 component score
view_score	Float	Signal 1 view fit score
audience_score	Float	Signal 5 audience demo score
cpm_score	Float	Signal 4 price/market fit score
score_version	String?	Scorer version tag (e.g. v3)
scored_at	DateTime?	When this score was computed

Total rows: ~107,000+ (as of last full V3 rescore).

→ See Matching & Scoring

---

brand_intel

One row per brand. Contains CRM data, scoring context, deal history, and AI-enriched intelligence.

Field	Type	Notes
id	Int	Primary key
name	String	Brand name (canonical form)
website	String?	Brand website from Close CRM
email	String?	Primary contact email from Close
avg_cpm	Float?	Weighted average CPM from matched creators
avg_rate	Float?	Average deal rate from monday_deliverables
total_deals	Int?	Total deal count from Monday
agent	String?	Assigned CA agent
last_activity_date	Date?	Most recent deal date (from Monday, synced 6-hourly)
last_contacted_date	DateTime?	Most recent outbound email timestamp (from Close, synced daily)
last_contact_agent	String?	Agent who sent the most recent email
last_contact_subject	String?	Subject of the most recent email
last_contact_summary	String?	AI-generated summary of the most recent email thread
profile	JSON?	Enriched brand profile: { creators, verticals, audience, content_themes, ideal_creator_archetype, platforms }
what_they_want	JSON?	Simplified targeting profile
fathom_calls	JSON?	Array of past Fathom call objects with summaries
deal_history	JSON?	Array of past deals with creator, value, close date
agencies	JSON?	Other agencies this brand works with

Total rows: 2,786 brands (1,178 original + 1,608 from Close import).

---

creators

Creator profile data synced from Airtable. Used for match context and the creator detail page.

Field	Type	Notes
id	Int	Primary key
name	String	Creator name (canonical, matches matches.creator)
slug	String	URL slug for the creator detail page
platform	String	Primary platform (youtube, tiktok, etc.)
subscribers	Int?	Subscriber/follower count
avg_views	Int?	Average views per video
rate	Int?	Rate in dollars
cpm	Float?	CPM value
niches	String?	Niche tags
headshot_url	String?	Airtable CDN URL for headshot
airtable_id	String?	Airtable record ID for this creator

Synced from Airtable on ca-data startup and every 2 hours via src/sync.js.

---

brand_emails

Latest email thread per brand, sourced from Close CRM. One row per brand (the most recent thread).

Field	Type	Notes
id	Int	Primary key
brand	String	Brand name
contact_emails	String[]	Email addresses involved in the thread
thread_id	String	Gmail thread ID
thread_subject	String	Email subject
thread_date	DateTime	Date of the most recent message
agent	String	CA agent whose inbox the thread was found in
summary	String?	AI-generated summary (gpt-4o-mini)

Synced by sync-close-contacts.js (daily, 4am UTC). Searches across 5 agent Gmail inboxes for threads matching each Close lead.

---

monday_deliverables

Deal records synced from Monday.com. Source of avg_rate, total_deals, and last_activity_date in brand_intel.

Field	Type	Notes
id	Int	Primary key
deal_name	String	Monday item name (e.g. "Caleb Hammer YouTube Integration for SoFi")
creator	String	Creator name (parsed + normalized from deal name)
brand	String	Brand name (parsed from deal name)
agent	String?	Agent column from Monday
creator_payment	Int?	Deal value in dollars
status	String?	Monday deal status
close_date	Date?	Deal close date
created_at	DateTime	Row creation time

Cancelled items are excluded from all rate/count aggregations.

Name normalization: Deal names are stripped of suffixes like (renewal), (extension), (upsell), (copy) before creator matching.

---

conflict_database

Active sponsor conflicts detected for roster creators. Used to show conflict warning badges on match cards.

Field	Type	Notes
id	Int	Primary key
creator	String	Creator name
brand	String	Brand the creator is actively sponsoring
detected_at	DateTime	When the conflict was first detected
source	String?	Detection source (e.g. video transcript ID)

Populated by the separate sponsor detection pipeline. ~39 creators have active detections.