launchkit-ai - GitLite

KO | EN

gitlite — search

#hacktoberfest #react #typescript #android #machine-learning #linux #docker #macos #nodejs #windows #vue #awesome-list #pytorch #ios

launchkit-ai

(★ 105)

Open-source AI app builder for turning prompts into runnable web apps with sandboxed coding agents.

파일 탐색기

최종 버전 다운로드 (.zip)

- - bug_report.md
  - feature_request.md
- - ci.yml
- PULL_REQUEST_TEMPLATE.md
- settings.json
- - .gitkeep
  - home-mobile.png
  - home.png
- demo.md
- release-v0.1.0.md
- roadmap-issues.md
- - migration_lock.toml
- schema.prisma
- file.svg
- globe.svg
- logo-text.svg
- logo.svg
- next.svg
- vercel.svg
- window.svg
- - compile_page.sh
  - e2b.Dockerfile
  - e2b.toml
- - error.tsx
  - favicon.ico
  - globals.css
  - layout.tsx
- - file-explorer.tsx
  - hint.tsx
  - tree-view.tsx
  - user-control.tsx
- - use-current-theme.ts
  - use-mobile.ts
  - use-scroll.ts
- - client.ts
  - functions.ts
  - types.ts
  - utils.ts
- - db.ts
  - usage.ts
  - utils.ts
- - client.tsx
  - init.ts
  - query-client.ts
  - server.tsx
- middleware.ts
- prompt.ts
- types.ts
.gitignore
CHANGELOG.md
components.json
CONTRIBUTING.md
eslint.config.mjs
LICENSE
next.config.ts
package-lock.json
package.json
postcss.config.mjs
README.md
ROADMAP.md
SECURITY.md
tsconfig.json

# 설치 가이드

1. 코드 내려받기

git clone https://github.com/AndyY-Q/launchkit-ai

깃허브에서 프로젝트 코드 전체를 내 컴퓨터로 내려받습니다.

cd launchkit-ai

방금 내려받은 프로젝트 폴더 안으로 이동합니다.

2. Node.js

쉬움 추천

사전 준비물

Git GitHub에서 프로젝트 코드를 내려받으려면 필요합니다.
Node.js npm 명령어를 쓰려면 Node.js가 설치되어 있어야 합니다. LTS(장기지원) 버전을 추천합니다.

npm install

package.json에 명시된 라이브러리들을 내려받아 설치합니다.

npm run dev

개발/실행 서버를 켭니다.

npm run lint

이 명령어를 터미널에 그대로 입력해 실행하세요.

npm run typecheck

이 명령어를 터미널에 그대로 입력해 실행하세요.

npm run build

배포용으로 코드를 최적화해서 빌드 파일을 생성합니다.

✅ 명령어 실행 후 터미널에 나타나는 주소(보통 http://localhost:3000 형태)를 브라우저에서 열어보세요.

이 레포의 README에 적힌 실제 명령어를 그대로 가져왔습니다.

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# LaunchKit AI

LaunchKit AI is an open-source AI app builder that turns prompts into runnable web apps with sandboxed coding agents. It combines a chat-first product surface with OpenAI-powered agent workflows, Inngest orchestration, E2B code sandboxes, tRPC APIs, Clerk authentication, Prisma, and Postgres.

The project is designed for maintainers and builders who want a practical reference app for prompt-to-app generation, safe sandbox execution, usage limits, and previewable artifacts.

## Why This Exists

Most AI app builders hide the interesting parts. LaunchKit AI keeps the architecture visible so open-source maintainers can inspect, fork, and adapt the core loop:

```mermaid
flowchart LR
  A["User prompt"] --> B["tRPC protected mutation"]
  B --> C["Inngest code-agent/run event"]
  C --> D["OpenAI agent network"]
  D --> E["E2B sandbox tools"]
  E --> F["Generated files + preview URL"]
  F --> G["Prisma project fragment"]
  G --> H["Next.js UI"]
```

## Features

- Sandboxed coding-agent workflow with `@inngest/agent-kit` and E2B.
- OpenAI model calls for code generation, fragment titles, and user-facing responses.
- Next.js App Router, React 19, Tailwind CSS, and shadcn-style UI primitives.
- tRPC routers for projects, messages, and usage state.
- Clerk authentication with protected project routes.
- Prisma schema for projects, messages, fragments, and usage accounting.
- Rate-limited free/pro usage model through `rate-limiter-flexible`.
- CI-ready lint, typecheck, and build scripts.

## Screenshots

![LaunchKit AI public home](https://raw.githubusercontent.com/AndyY-Q/launchkit-ai/main/docs/screenshots/home.png)

See [docs/demo.md](./docs/demo.md) for smoke-test verification and the full provider-backed demo checklist.

## Tech Stack

| Area | Tools |
| --- | --- |
| App | Next.js 15, React 19, TypeScript |
| UI | Tailwind CSS 4, Radix UI, lucide-react |
| API | tRPC, TanStack Query |
| Agents | OpenAI, Inngest Agent Kit |
| Sandboxes | E2B Code Interpreter |
| Auth | Clerk |
| Data | Prisma, Postgres |
| Background jobs | Inngest |

## Quick Start

```bash
git clone https://github.com/AndyY-Q/launchkit-ai.git launchkit-ai
cd launchkit-ai
npm install
cp .env.example .env.local
npm run dev
```

Open [http://localhost:3000](http://localhost:3000).

## Environment Variables

Create `.env.local` from `.env.example`.

| Variable | Required | Purpose |
| --- | --- | --- |
| `DATABASE_URL` | Yes | Postgres connection string used by Prisma. |
| `NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY` | Yes | Public Clerk browser key. CI can use the documented dummy test key for build-only checks. |
| `CLERK_SECRET_KEY` | Yes for auth runtime | Server-side Clerk key. |
| `OPENAI_API_KEY` | Yes for agent runtime | OpenAI key used by `@inngest/agent-kit`. |
| `E2B_API_KEY` | Yes for sandbox runtime | E2B key used to create sandboxes. |
| `INNGEST_EVENT_KEY` | Yes for deployed jobs | Inngest event key. |
| `INNGEST_SIGNING_KEY` | Yes for deployed jobs | Inngest signing key. |
| `NEXT_PUBLIC_APP_URL` | Recommended | Canonical app URL for deployed environments. |

## Provider Setup

### Clerk

1. Create a Clerk application.
2. Copy the publishable and secret keys into `.env.local`.
3. Configure sign-in and sign-up URLs for `/sign-in` and `/sign-up`.

For CI build verification only, this repository uses a non-secret dummy publishable key:

```bash
NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY=pk_test_ZHVtbXkuY2xlcmsuYWNjb3VudHMuZGV2JA
CLERK_SECRET_KEY=sk_test_ZHVtbXk
```

These values are intentionally not valid for real runtime authentication.
When these exact dummy values are present, LaunchKit AI enters smoke-test mode: Clerk middleware, ClerkJS, the pricing table, and authenticated project lists are skipped so CI can verify the public app shell without contacting Clerk.

### OpenAI

Create an API key and set `OPENAI_API_KEY`. The current agent configuration lives in `src/inngest/functions.ts`.

### E2B

Create an E2B account and set `E2B_API_KEY`. The sandbox template is in `sandbox-templates/nextjs`.

### Inngest

Run Inngest locally when testing background jobs:

```bash
npx inngest-cli@latest dev
```

The app exposes the Inngest handler at `/api/inngest`.

### Database

Run migrations after configuring `DATABASE_URL`:

```bash
npx prisma migrate dev
```

## Development

```bash
npm run dev
npm run lint
npm run typecheck
npm run build
```

The `npm run ci` command runs all three verification steps.

## Repository Visibility Checklist

Before applying to OpenAI Codex for OSS, make sure GitHub can review the project without authentication:

- Set the repository visibility to public.
- Confirm `https://github.com/AndyY-Q/launchkit-ai` loads in a private browser window.
- Confirm the repository appears on the public `AndyY-Q` profile.
- Confirm the GitHub API returns metadata from `https://api.github.com/repos/AndyY-Q/launchkit-ai`.
- Add repository topics such as `openai`, `coding-agent`, `e2b`, `inngest`, `nextjs`, `trpc`, `oss`.
- Add a short repository description: `Open-source AI app builder for turning prompts into runnable web apps with sandboxed coding agents.`

## OSS Application Narrative

LaunchKit AI should be presented as an open-source AI app builder with a transparent sandboxed coding-agent architecture. The strongest application angle is:

> LaunchKit AI gives open-source maintainers and builders a working reference for prompt-to-app generation, isolated code execution, artifact previews, and usage governance using widely adopted web primitives.

## Contributing

Contributions are welcome. Start with issues labeled `good first issue` or `help wanted`, and open a pull request with a clear description, screenshots for UI changes, and the output of:

```bash
npm run ci
```

## Roadmap

See [ROADMAP.md](./ROADMAP.md).

## License

MIT. See [LICENSE](./LICENSE).