prisma-cli

Run prisma migrate dev with an optional --name flag to create and apply a migration from your schema changes. It uses a shadow database for drift detection, and you should run prisma generate explicitly afterward when you need refreshed client output.

db push syncs the schema directly without creating migration files or history, making it suited for prototyping and MongoDB. migrate dev creates tracked migration files with rollback capability and is the standard for team development workflows.

Use bunx --bun before Prisma commands, such as bunx --bun prisma init or bunx --bun prisma generate. The --bun flag ensures Prisma runs with the Bun runtime instead of falling back to Node.js because of the CLI shebang.

Prisma detects AI agent environments and blocks destructive commands like migrate reset and db push --force-reset until the user gives explicit consent. The agent must explain the data-loss impact, obtain consent immediately before running, and pass the exact consent text via PRISMA_USER_CONSENT_FOR_DANGEROUS_AI_ACTION.

No, in current Prisma versions you must run prisma generate and prisma db seed explicitly after migrate dev when you need refreshed client output or seed data. The seed script is configured in prisma.config.ts under the migrations section.

Use migrate deploy in production, staging, and CI/CD pipelines because it only applies pending migrations without prompts, drift detection, or a shadow database. Never run migrate dev in production environments.