LangGraph chat architecture

Automatically synced with your v0.app deployments

Why I built this

Not AI as a chat bubble, but as the thing that drives the UI. An expense tracker as the testbed:

• talk in plain language → the agent picks the tools and renders the widgets 📊
• every write goes through an explicit changeset review — no silent edits 🔁
• the point isn't "chatbot for money," it's the interface pattern

Overview

This repository will stay in sync with your deployed chats on v0.app. Any changes you make to your deployed app will be automatically pushed to this repository from v0.app.

Backend modes (mock vs. real)

The chat assistant calls a backend API for all financial data. To keep the app fully functional without standing up the ai-expense-tracker-service, it ships with a built-in, in-memory mock backend (lib/api/mock/) that returns realistic data in the same shapes as the real service.

• Mock mode (default): every API call is served from memory, so the demo works on Vercel with zero infrastructure. Reads are always consistent; writes persist for the life of a warm serverless instance (see the note in lib/api/mock/router.ts).
• Real mode: set NEXT_PUBLIC_USE_MOCK_API=false and point NEXT_PUBLIC_API_BASE_URL at a deployed backend.

Variable	Default	Purpose
NEXT_PUBLIC_USE_MOCK_API	true	true/unset → mock; false → real backend
NEXT_PUBLIC_API_BASE_URL	—	Real backend URL (used only when mock is off)
OPENAI_API_KEY	—	Server-side LLM access for the chat route

NEXT_PUBLIC_* values are inlined at build time — change them in Vercel and redeploy for the change to take effect. See .env.example.

Deployment

Your project is live at:

https://vercel.com/kimsing-ais-projects/v0-lang-graph-chat-architecture

Build your app

Continue building your app on:

https://v0.app/chat/rLUz501f5R7

How It Works

1. Create and modify your project using v0.app
2. Deploy your chats from the v0 interface
3. Changes are automatically pushed to this repository
4. Vercel deploys the latest version from this repository