Usage

Try it in 2 minutes

Run the NestJS/Prisma preset into a throwaway folder:

npx backend-ai-starter-recipes --preset nestjs-prisma --output ./my-backend

Then inspect the generated operating manual:

my-backend/
├── .ai/
│   ├── AGENT.md
│   ├── lifecycle/
│   ├── rules/
│   ├── skills/
│   ├── context/
│   └── tracking/
└── .cursor/ or other IDE adapter files

Use bare --preset nestjs-prisma --output ./my-backend after a global install.

Two modes

Mode	When to use
Interactive	First time, or a custom stack not covered by a preset
--preset	CI, docs, or a known stack (Nest + Prisma, etc.)

Interactive flow (grouped)

The CLI walks through project identity → stack → tooling → outputs. Roughly:

What you will be asked

• Framework: NestJS, Express, Fastify, Hono
• Language: TypeScript or JavaScript
• ORM: Prisma, TypeORM, Drizzle, MikroORM, Knex, or none (raw SQL)
• Database: PostgreSQL, MySQL, MongoDB, SQLite
• Validation: class-validator, Zod, Joi, or none
• API docs: Swagger/OpenAPI or none
• Auth: JWT (Passport), session, OAuth2 provider, custom, or none
• Tests: Jest, Vitest, Mocha
• Monitoring: Sentry, APM, Prometheus (multi-select)
• CI/CD: GitHub Actions, GitLab CI, or none
• IDEs: Cursor, Claude Code, VS Code Copilot, Antigravity, Windsurf, or all
• Skills: plan-review, code-review, QA, ship, plus optional workflows (document-release, retro, db-migration-review, api-contract-check, dependency-audit) or all

If you omit --output, the CLI prompts for a directory (and warns if it is non-empty).

Preset coverage

Preset	Stack (summary)
nestjs-prisma	NestJS, TypeScript, Prisma, PostgreSQL, class-validator, Swagger/OpenAPI, JWT, Jest
nestjs-typeorm	NestJS, TypeScript, TypeORM, PostgreSQL, class-validator, Swagger/OpenAPI, JWT, Jest
express-prisma	Express, TypeScript, Prisma, PostgreSQL, Zod, Swagger/OpenAPI, JWT, Vitest
fastify-drizzle	Fastify, TypeScript, Drizzle, PostgreSQL, Zod, Swagger/OpenAPI, JWT, Vitest

nestjs-prisma

npx backend-ai-starter-recipes --preset nestjs-prisma --output ./my-nestjs-app

nestjs-typeorm

npx backend-ai-starter-recipes --preset nestjs-typeorm --output ./my-app

express-prisma

npx backend-ai-starter-recipes --preset express-prisma --output ./my-express-app

fastify-drizzle

npx backend-ai-starter-recipes --preset fastify-drizzle --output ./my-fastify-app

CLI flags

Flag	Short	Description
--output <dir>	-o	Output directory (skips the path prompt)
--preset <name>	-p	Use a JSON preset from the package’s presets/ folder

Minimal example

npx backend-ai-starter-recipes --preset nestjs-prisma --output ./api

You should see a tree similar to:

api/
├── .ai/
│   ├── AGENT.md
│   ├── rules/
│   ├── skills/
│   ├── context/
│   └── tracking/
├── .cursor/          # if Cursor was selected in preset
└── ...

(Preset defaults include specific IDEs; customize via interactive run or by editing a copied preset JSON for your fork.)

Known Limitations

• This is an early community release intended for developer testing and feedback.
• Presets are opinionated starting points, not proof that every team using that stack should follow the same rules.
• Generated .ai/ content should be reviewed and edited inside your real repo before treating it as authoritative.
• IDE adapters depend on how each AI tool reads repository context; behavior may differ across tool versions.
• The CLI creates AI instructions, lifecycle guidance, rules, skills, context, tracking, and adapter files. It does not scaffold a complete backend service.

---

Next: learn what each folder means — Understanding the output.