Stack Tecnológica

Stack Tecnológica

Stack choices optimize legal data integrity first

The selected stack emphasizes reliable schema evolution, explicit typing, and compatibility with both human-facing and AI-agent-facing layers.

Runtime and language stack

Core runtime components:

• Node.js >=20.0.0
• npm >=10.0.0
• TypeScript (strict mode)
• Next.js 16 + React 19

Primary dependencies and roles

Dependency	Role in project
next	Web runtime and routing shell
react, react-dom	UI rendering
drizzle-orm	Type-safe schema modeling
pg	PostgreSQL client for scripts/runtime integration
drizzle-kit	Schema/migration tooling support
tsx	TypeScript script execution (migrations and tooling)
eslint, eslint-config-next	Linting baseline
tailwindcss, @tailwindcss/postcss	Styling pipeline

Database and indexing stack

Database layer uses PostgreSQL with pgvector enabled. Search and retrieval foundations include:

• generated tsvector index for Portuguese full-text search;
• IVFFlat vector index for embedding similarity workflows.

Build and quality toolchain

Key commands from package.json:

• npm run dev
• npm run build
• npm run start
• npm run lint
• npm test

External services currently referenced

• Railway PostgreSQL (connection through DATABASE_URL)
• GitHub Actions workflow using DISCORD_PR_REVIEW_WEBHOOK

Upgrade posture

Dependency updates should be staged by risk:

1. patch upgrades for low-risk runtime/tooling;
2. minor upgrades after lint/build/test validation;
3. major upgrades only with explicit compatibility testing.

Cuidado

Keep migration and schema tooling compatible before bumping DB-related dependencies. Stack upgrades that break migration flow have high operational impact.