DEV Community

Cover image for Build Your First AI Agent with NestJS + RapidKit in 10 Minutes
RapidKit
RapidKit

Posted on

Build Your First AI Agent with NestJS + RapidKit in 10 Minutes

#rapidkit #ai #agents #nestjs

Adding AI features to NestJS usually takes much longer than it should.

With RapidKit, you can scaffold a production-style AI assistant module in minutes.

What You Build

  • AI assistant endpoints in NestJS
  • Provider abstraction (echo by default)
  • Optional OpenAI provider via environment
  • Completion + stream APIs
  • Support ticket endpoint with urgency classification
  • Health + Swagger visibility

Source code: https://github.com/getrapidkit/rapidkit-examples/tree/main/my-ai-workspace/ai-agent-nest

Prerequisites

  • Node.js 20+
  • npm 10+
  • Python 3.10+

1) Create workspace and project 

npx rapidkit my-ai-workspace
cd my-ai-workspace
rapidkit create project nestjs.standard ai-agent-nest
cd ai-agent-nest
rapidkit init
Enter fullscreen mode Exit fullscreen mode

2) Add AI Assistant module 

npx rapidkit add module ai_assistant
Enter fullscreen mode Exit fullscreen mode

This generates:

  • src/modules/free/ai/ai_assistant/ai_assistant.module.ts
  • src/modules/free/ai/ai_assistant/ai_assistant.controller.ts
  • src/modules/free/ai/ai_assistant/ai_assistant.service.ts
  • src/modules/free/ai/ai_assistant/ai_assistant.validation.ts

3) Register module in src/modules/index.ts

For latest ai_assistant module versions this is injected automatically.
If you use an older generated project and routes are missing, add this entry to optionalModules:

registerOptionalModule(() => require('./free/ai/ai_assistant').AiAssistantModule as ModuleRef),
Enter fullscreen mode Exit fullscreen mode

4) Run app 

PORT=8013 npm run start:dev
Enter fullscreen mode Exit fullscreen mode

or via RapidKit wrapper:

rapidkit dev -p 8013
Enter fullscreen mode Exit fullscreen mode

To auto-register OpenAI provider:

OPENAI_API_KEY=your_key PORT=8013 npm run start:dev
Enter fullscreen mode Exit fullscreen mode

5) Verify endpoints 

curl http://127.0.0.1:8013/health
curl http://127.0.0.1:8013/ai/assistant/providers
curl http://127.0.0.1:8013/ai/assistant/health
Enter fullscreen mode Exit fullscreen mode

Providers response:

["echo"]
Enter fullscreen mode Exit fullscreen mode

Completion test:

curl -X POST http://127.0.0.1:8013/ai/assistant/completions \
  -H "Content-Type: application/json" \
  -d '{"prompt":"What is RapidKit?","provider":"echo"}'
Enter fullscreen mode Exit fullscreen mode

Example response:

{
  "provider": "echo",
  "content": "[echo] What is RapidKit?",
  "latencyMs": 0,
  "cached": false
}
Enter fullscreen mode Exit fullscreen mode

Stream test:

curl -X POST http://127.0.0.1:8013/ai/assistant/stream \
  -H "Content-Type: application/json" \
  -d '{"prompt":"Give me a deployment checklist","provider":"echo"}'
Enter fullscreen mode Exit fullscreen mode

Cache clear:

curl -X DELETE http://127.0.0.1:8013/ai/assistant/cache
Enter fullscreen mode Exit fullscreen mode

Support ticket parity endpoint:

curl -X POST http://127.0.0.1:8013/support/ticket \
  -H "Content-Type: application/json" \
  -d '{"message":"Customer cannot complete payment"}'
Enter fullscreen mode Exit fullscreen mode

Example response:

{
  "ticket_id": "TKT-12345",
  "urgency": "high",
  "ai_response": "[echo] Customer cannot complete payment",
  "latency_ms": 0,
  "next_action": "escalate",
  "provider": "echo"
}
Enter fullscreen mode Exit fullscreen mode

6) Open API docs

  • http://127.0.0.1:8013/docs

Why this setup is useful

  • Fast baseline for real backend teams
  • Clear module boundaries for scaling
  • Easy to swap provider strategy later

FastAPI parity now exists in this example for the support-ticket flow and optional OpenAI path.

If you want, next step is adding OpenAI provider + retry/fallback in this same NestJS codebase.

Top comments (0)