chat-api

Express chat API with Supabase (conversations + messages + system prompt table), OpenAI, and rate limiting. Mount in your app or run standalone. Use with chat on the frontend.

Install

npm install chat-api
# or from Git
npm install git+ssh://[email protected]:saulommartins/chat.api.git

Database (Supabase)

Run the migrations in your Supabase SQL Editor (in order):

1. migrations/01-conversations-messages.sql — creates conversations and messages
2. migrations/02-chat_system_prompt.sql — creates chat_system_prompt (single row, id=1). Edit the row content to set your assistant behavior and optional WhatsApp link.

Mount in your Express app

const express = require('express');
const { createClient } = require('@supabase/supabase-js');
const { createChatRouter } = require('chat-api');

const supabase = createClient(
  process.env.SUPABASE_URL,
  process.env.SUPABASE_SERVICE_KEY
);

const chatRouter = createChatRouter({
  supabase,
  openaiApiKey: process.env.OPENAI_API_KEY,
  rateLimitWindowMs: Number(process.env.RATE_LIMIT_CHAT_WINDOW_MS) || 60000,
  rateLimitMax: Number(process.env.RATE_LIMIT_CHAT_MAX) || 10,
  // optional: rateLimitMessage, conversationsTable, messagesTable, systemPromptTable, openaiModel, openaiMaxTokens, openaiTemperature
});

const app = express();
app.use(express.json());
app.use('/api/chat', chatRouter);

Options (ChatApiOptions)

| Option | Type | Default | Description | |--------|------|---------|-------------| | supabase | SupabaseClient | required | Supabase client (service role). | | openaiApiKey | string | required | OpenAI API key. | | rateLimitWindowMs | number | 60000 | Rate limit window (ms). | | rateLimitMax | number | 10 | Max POST requests per window per IP. | | rateLimitMessage | string | (en) | JSON message when limit exceeded. | | conversationsTable | string | conversations | Table name. | | messagesTable | string | messages | Table name. | | systemPromptTable | string | chat_system_prompt | Table name (single row, id=1). | | openaiModel | string | gpt-3.5-turbo | OpenAI model. | | openaiMaxTokens | number | 500 | Max tokens per reply. | | openaiTemperature | number | 0.7 | Temperature. |

API contract

• POST /api/chat — Body: { message: string, conversationId?: string }. Returns { conversationId, reply, message }. Rate limited.
• GET /api/chat/:conversationId — Returns { conversationId, messages }.

Repo

https://github.com/saulommartins/chat.api