https://github.com/mohatt/chatbot-starter
Full‑stack AI chatbot starter kit.
https://github.com/mohatt/chatbot-starter
ai ai-sdk chatbot nextjs vercel
Last synced: 4 months ago
JSON representation
Full‑stack AI chatbot starter kit.
- Host: GitHub
- URL: https://github.com/mohatt/chatbot-starter
- Owner: mohatt
- License: apache-2.0
- Created: 2026-02-03T08:19:54.000Z (5 months ago)
- Default Branch: main
- Last Pushed: 2026-03-09T16:25:41.000Z (4 months ago)
- Last Synced: 2026-03-09T16:49:49.346Z (4 months ago)
- Topics: ai, ai-sdk, chatbot, nextjs, vercel
- Language: TypeScript
- Homepage: https://chat.mohatt.dev
- Size: 2.08 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Chatbot Starter
Full‑stack AI chatbot starter with RAG, file uploads, and project‑scoped chats, built on **Next.js**, **Vercel AI SDK**, **Postgres**, and **Upstash Vector**.
[Live Demo][vercel-demo]
## Features
- **Streaming chat UI** with Markdown rendering, auto-scroll, and rich AI elements.
- **Multi-model support** via Vercel AI Gateway (OpenAI, Anthropic, Google, xAI, and many more).
- **File uploads + RAG** with vector indexing and in-chat file search/read tools.
- **Conversation branching** with immutable message trees.
- **Per-message branch navigation** (`1 of N`) to switch between message variants inline.
- **Web search** support for OpenAI, Anthropic, and Google Gemini models.
- **Projects + chats** with per-project history and organized sidebar.
- **Auth + access control** with public/private chat visibility.
- **System emails** for user account management.
- **Usage tracking & limits** with per-user chat credits.
- **Advanced routing + client-side caching** for seamless navigation and performance.
## Tech Stack
- **Backend:**
- **AI layer:** Vercel AI SDK + AI Gateway
- **DB layer:** Neon Postgres + Drizzle ORM
- **RAG:** Upstash Vector
- **File storage:** Vercel Blob
- **Authentication:** Better Auth
- **Emails:** Resend API
- **Frontend:**
- Next.js 16 + React 19 + React Query
- shadcn/ui + Tailwind CSS
- **Hosting:**
- Vercel Platform
## Deploy Your Own
You can deploy your own version of the Chatbot to Vercel with one click:
[][vercel-deploy]
## Setup
### 1. Install dependencies
```bash
pnpm install
```
### 2. Database
The Vercel dashboard will guide you through provisioning **Neon Postgres** and **Upstash Vector** (the required integrations) during setup.
Once connected, Vercel will inject the required environment variables automatically.
### 3. Environment variables
Most environment variables are provisioned by Vercel during setup and can be pulled locally with:
```bash
# Install Vercel CLI
pnpm i -g vercel
# Link local instance with Vercel and GitHub accounts
vercel link
# Download your environment variables locally
vercel env pull .env.development.local
```
You only need to set the required secrets (e.g., `AUTH_SECRET`, `CRON_SECRET`). You may use `openssl rand -base64 32` to generate those secrets.
Refer to [`.env.example`](.env.example) for more information.
> **Recommended:** Set env vars in the Vercel dashboard so they’re automatically included whenever you pull envs.
#### Enable system emails for user account management (Optional)
To enable user account emails (email verification and password reset), configure:
```bash
RESEND_API_KEY="abc-123"
EMAIL_SENDER_ADDRESS="no-reply@yourdomain.com"
```
### 4. Apply database migrations
After your Postgres URL is set locally, run:
```bash
pnpm db:push
```
This will sync the current schema in [`lib/db/schema`](./lib/db/schema) to your Postgres database.
### 5. Review app configurations
Most app configurations live in [`lib/config.ts`](./lib/config.ts), including:
- AI models registry and defaults
- Chat limits (message parts, history, max messages per chat)
- File upload rules and size limits
- Billing tiers and usage caps
- Project limits (e.g., max files)
### 6. Run locally
```bash
pnpm dev
```
Visit `http://localhost:3000`.
### 7. Deploy to Vercel
Once the setup steps above are complete, deploy from the Vercel dashboard:
1. Open your project in Vercel.
2. Click **Deploy** (or push to your connected Git branch).
Or deploy from the CLI:
```bash
vercel deploy
```
[vercel-demo]: https://chat.mohatt.dev
[vercel-deploy]: https://vercel.com/new/clone?repository-url=https%3A%2F%2Fgithub.com%2Fmohatt%2Fchatbot-starter&env=AUTH_SECRET,CRON_SECRET&envDescription=API%20Keys%20needed%20for%20the%20application.&envLink=https%3A%2F%2Fgithub.com%2Fmohatt%2Fchatbot-starter%23env-vars&project-name=ai-chatbot&repository-name=ai-chatbot&demo-title=Chatbot%20Starter&demo-description=A%20full-stack%20AI%20Chatbot%20Starter%20powered%20by%20Next.js%20and%20Vercel%20AI%20SDK.&demo-url=https%3A%2F%2Fchatbot-starter.vercel.app&demo-image=https%3A%2F%2Fraw.githubusercontent.com%2Fmohatt%2Fchatbot-starter%2Frefs%2Fheads%2Fmain%2Fpublic%2Fproject-logo.png&stores=%5B%7B%22type%22%3A%22integration%22%2C%22protocol%22%3A%22storage%22%2C%22productSlug%22%3A%22neon%22%2C%22integrationSlug%22%3A%22neon%22%7D%2C%7B%22type%22%3A%22integration%22%2C%22protocol%22%3A%22storage%22%2C%22productSlug%22%3A%22upstash-vector%22%2C%22integrationSlug%22%3A%22upstash%22%7D%2C%7B%22type%22%3A%22blob%22%7D%5D