🚀 LaunchSmart

LaunchSmart is a full-stack, AI-powered pre-launch QA platform built on Next.js and Supabase. It provides engineering and product teams with a single command center to definitively know whether their product is ready to ship.

By leveraging the multi-modal capabilities of Gemini 2.5 Flash, LaunchSmart acts as a senior release engineer by statically analyzing code diffs, verifying visual UI implementation against specs, identifying required fixes, and directly drafting highly structured bug tickets to Jira or GitHub.

---

✨ Core Features

👁️ AI Vision QA Pipeline

Describe what you're shipping: type it out, paste it, upload a file, or link a Jira ticket. Then show us the build: upload UI screenshots, drop in a GitHub PR link, or both. The pipeline:

1. Extracts expected behaviors from the product spec text.
2. Runs multimodal vision analysis on the actual UI screenshots.
3. Compares the two to find gaps, missing states, or incorrect flows.
4. Outputs up to 5 structured issues, each containing severity, root cause, recommended fix, and confidence level.
5. Computes a strict, deterministic 0-100 launch score (No AI inflation: -25 per critical, -10 per high, -4 per medium issue). Buckets the release into Ready, Launch with Caution, or Block Release.

💻 Autonomous Agentic PR Checker (Powered by MCP)

Paste a GitHub PR URL and optionally link a Jira ticket. LaunchSmart unleashes Gemini as an autonomous agent:

1. Built-in Model Context Protocol (MCP) Server securely exposes API integrations as executable tools.
2. The AI explicitly determines exactly when and what data to fetch (e.g., triggering fetch_pr_diff and fetch_jira_ticket) mid-thought loop.
3. Dynamically interprets the code diffs and Jira Acceptance criteria iteratively rather than being constrained to rigid, massive static prompts.
4. Return an APPROVE or BLOCK verdict with a detailed breakdown of what is implemented vs. missing in the code.

🔗 Seamless Integrations

Push findings to your existing workflows with one click:

• Jira: Generates fully formatted bug tickets using Atlassian Document Format (ADF), including actual behavior, expected behavior, recommended fixes, and acceptance checks. Built-in JQL duplicate detection ensures you don't file the same bug twice.
• GitHub Issues: Generates Markdown-formatted issues with a launchsmart label, also preventing duplicates.
• Slack: Sends a rich Block Kit card summary to your team's channel containing the score, verdict, issue breakdown, and a deep link back to the report.

🔒 Bring Your Own Key (BYOK) Integration

• User-Specific Keys: Users can navigate to the /settings page to securely store their own individual GitHub Personal Access Tokens and Jira API credentials.
• Secure Storage: Keys are saved to the user-scoped user_api_keys Supabase table and fetched on demand.
• Seamless Fallback: All API routes (PR Check, Ticket Generation, etc.) intelligently check for a user's stored keys first and smoothly fallback to instance-wide environment variables if not present.

💾 Relational Persistence & Fallbacks

• All reports, findings, PR checks, and created ticket records are persisted to Supabase.
• User-scoped auth relies on GitHub OAuth and built-in Supabase auth.
• Built-in localStorage fallback ensures the application remains functional even if the database is temporarily unreachable.

---

🛠️ Tech Stack

• Frontend: Next.js (App Router), React, Tailwind CSS, shadcn/ui, Zustand
• Backend: Next.js Serverless API Routes
• AI / LLM: @ai-sdk/google (Gemini 2.5 Flash), @modelcontextprotocol/sdk (MCP Agentic Loop)
• Database & Auth: Supabase
• Integrations: GitHub API, Atlassian (Jira) REST API v3, Slack Webhooks

---

🚀 Getting Started

1. Clone the repository

git clone https://github.com/yourusername/launchsmart.git
cd launchsmart

2. Install dependencies

npm install
# or
pnpm install
# or
yarn install

3. Setup Environment Variables

Create a .env.local file in the root directory and add the following keys:

# AI Model
GOOGLE_GENERATIVE_AI_API_KEY=your_gemini_api_key

# Supabase
NEXT_PUBLIC_SUPABASE_URL=your_supabase_project_url
NEXT_PUBLIC_SUPABASE_ANON_KEY=your_supabase_anon_key

# Integrations (Optional: Instance-Wide Fallbacks)
# If users do not provide their own keys via the /settings page,
# the application will fall back to these global variables.
GITHUB_TOKEN=your_github_personal_access_token
GITHUB_REPO=owner/repo # Default fallback repo for issues

JIRA_EMAIL=your_jira_email
JIRA_API_TOKEN=your_jira_api_token
JIRA_DOMAIN=your_company.atlassian.net

SLACK_WEBHOOK_URL=your_slack_incoming_webhook_url

4. Setup the Database Schema

Be sure to push the Supabase migrations to create the required tables, including the new user_api_keys table for BYOK support:

npx supabase db push

5. Run the Development Server

npm run dev
# or 
pnpm dev

Open http://localhost:3000 in your browser to see the application.

---

📚 How the Scoring Works

The Launch Score is deterministic to prevent LLM hallucination or overly optimistic grading:

• Base Score: Starts at 100
• Critical Issue: -25 points
• High Issue: -10 points
• Medium Issue: -4 points
• Unverified Critical: -6 points (if a critical piece of the spec is missing from the screenshots entirely)

Buckets:

• 90 - 100: Ready
• 70 - 89: Launch with Caution
• 0 - 69: Block Release

---

🤝 Contributing

Contributions, issues, and feature requests are welcome!