Get Started- Complete documentation overview

System prompts and prompt templates help you customize AI behavior and maintain consistent responses.

Overview

The prompt system (src/lib/ai/prompts.ts) provides:

  • System prompt templates - Pre-built prompts for common use cases
  • Prompt builders - Utilities to construct custom prompts
  • Message history management - Build conversation context
  • Prompt validation - Ensure prompts meet safety requirements

Default System Prompts

The boilerplate includes several pre-built system prompts:

General Assistant

import { DEFAULT_SYSTEM_PROMPT } from "@/lib/ai/prompts";

// Default: Helpful, friendly, professional assistant

Code Assistant

import { CODE_ASSISTANT_PROMPT } from "@/lib/ai/prompts";

// Optimized for coding tasks, debugging, explanations

Creative Writer

import { CREATIVE_WRITER_PROMPT } from "@/lib/ai/prompts";

// Optimized for creative writing, content improvement

Tiered System Prompts

The boilerplate includes tiered prompts based on subscription status:

Basic Tier Prompt

Used for free tier users (if free tier is enabled):

import { BASIC_TIER_PROMPT, getTierPrompt } from "@/lib/ai/prompts";

// Friendly, approachable, fundamental concepts
const prompt = getTierPrompt("free");

Characteristics:

  • Friendly and approachable tone
  • Fundamental concepts explained simply
  • Practical, actionable advice
  • Step-by-step guidance

Advanced Tier Prompt

Used for premium subscribers:

import { ADVANCED_TIER_PROMPT, getTierPrompt } from "@/lib/ai/prompts";

// Professional, technical, comprehensive
const prompt = getTierPrompt("premium");

Characteristics:

  • Professional and technical tone
  • In-depth technical analysis
  • Advanced concepts and methodologies
  • Comprehensive solutions

Using Tiered Prompts

import { buildMessageHistoryWithTier } from "@/lib/ai/prompts";
import { getSubscriptionTier } from "@/features/ai/server";

// Get user's tier
const tier = await getSubscriptionTier(userId);

// Build message history with tier-appropriate prompt
const messages = buildMessageHistoryWithTier(
  [{ role: "user", content: "Hello" }],
  { tier }
);

Building Custom Prompts

Using buildSystemPrompt()

import { buildSystemPrompt } from "@/lib/ai/prompts";

const customPrompt = buildSystemPrompt({
  systemPrompt: "You are a helpful assistant.",
  context: "User is working on a React project.",
  instructions: [
    "Always provide code examples",
    "Explain your reasoning",
    "Suggest best practices"
  ],
  examples: [
    {
      user: "How do I use hooks?",
      assistant: "React hooks let you use state..."
    }
  ]
});

Parameters

  • systemPrompt - Base system prompt
  • context - Additional context about the user or task
  • instructions - Array of specific instructions
  • examples - Array of example conversations

Message History

Building Message History (Legacy)

import { buildMessageHistory } from "@/lib/ai/prompts";

const messages = buildMessageHistory(
  [
    { role: "user", content: "Hello" },
    { role: "assistant", content: "Hi there!" }
  ],
  "You are a helpful assistant."
);

// Result includes system prompt at the beginning

Building Message History with Tier (Recommended)

import { buildMessageHistoryWithTier } from "@/lib/ai/prompts";
import { getSubscriptionTier } from "@/features/ai/server";

// Get user's subscription tier
const tier = await getSubscriptionTier(userId);

// Build with tier-appropriate prompt
const messages = buildMessageHistoryWithTier(
  [
    { role: "user", content: "Hello" },
    { role: "assistant", content: "Hi there!" }
  ],
  { tier } // Automatically uses BASIC_TIER_PROMPT or ADVANCED_TIER_PROMPT
);

// Or use custom system prompt
const messages = buildMessageHistoryWithTier(
  [{ role: "user", content: "Hello" }],
  { systemPrompt: "You are a custom assistant." }
);

Message Format

interface ChatMessage {
  role: "user" | "assistant" | "system";
  content: string;
}

Prompt Validation

Validate Prompts

import { validatePrompt } from "@/lib/ai/prompts";

const validation = validatePrompt([
  { role: "system", content: "..." },
  { role: "user", content: "..." }
]);

if (!validation.valid) {
  console.error(validation.error);
}

Validation checks:

  • Maximum message length
  • Maximum conversation length
  • Prohibited content patterns
  • Injection attempt detection

Examples

Custom AI Assistant

const assistantPrompt = buildSystemPrompt({
  systemPrompt: "You are a professional business consultant.",
  context: "The user is a startup founder.",
  instructions: [
    "Provide actionable advice",
    "Be concise and direct",
    "Use business terminology"
  ]
});

const response = await generateChatCompletion({
  messages: buildMessageHistory(
    [{ role: "user", content: "How do I raise funding?" }],
    assistantPrompt
  )
});

Domain-Specific Prompts

// Medical assistant (with disclaimers)
const medicalPrompt = buildSystemPrompt({
  systemPrompt: "You are a medical information assistant.",
  instructions: [
    "Always include disclaimers",
    "Do not provide diagnoses",
    "Recommend consulting healthcare professionals"
  ]
});

// Legal assistant
const legalPrompt = buildSystemPrompt({
  systemPrompt: "You are a legal information assistant.",
  instructions: [
    "Provide general information only",
    "Always recommend consulting an attorney",
    "Do not provide specific legal advice"
  ]
});

Best Practices

  1. Be specific - Clear instructions produce better results
  2. Include examples - Show the AI what you want
  3. Set boundaries - Define what the AI should and shouldn't do
  4. Test prompts - Iterate to find what works best
  5. Validate inputs - Always validate user prompts before sending

Security Considerations

  • Never trust user input - Always validate and sanitize
  • Use moderation - Check prompts for harmful content
  • Limit length - Prevent prompt injection attacks
  • Monitor usage - Track prompt patterns for abuse

Next Steps