Data Model

The primary v2 datastore is Supabase/PostgreSQL.

Supabase configuration

• Config: supabase/config.toml
• Migrations: supabase/migrations
• Rollbacks: supabase/migrations/rollbacks
• Canonical schema dump: backend/core-api/db_schema.sql
• Executable v2 baseline: supabase/migrations/20260304010000_v2_executable_schema.sql

Local Supabase ports from config:

• API: 54321
• DB: 54322
• Studio: 54323
• Inbucket: 54324
• DB major version: 17

Major table families

Domain	Tables
Identity/RBAC	profiles, roles, role_permissions, user_platform_roles, api_keys, api_key_logs
Teams/groups/projects	teams, team_members, team_roles, groups, group_members, circles, projects
Work	tasks, task_assignees, task_reviewers, task_applications, milestones, documents, resources
Finance	tokens, payment_token_groups, payments, transactions, transaction_legs, project_funding, treasury tables
Escrow/web3	chains, escrows, funding events, proposed operations, operation signatures, async jobs
Notifications/integrations	notifications, notification_events, preferences/templates, registered webhooks, delivery logs
Community/analytics	feed, reputation, Matrix space/room tables

Query guidelines

From AGENTS.md:

• Prefer database filtering/aggregation over Python post-processing.
• Select only needed columns.
• Avoid select("*") unless necessary.
• Use exact counts where needed.
• Use RPC for complex or repeated aggregation logic.