Simpli.Jira

An AI-powered system that autonomously evaluates whether a GitHub Pull Request satisfies the requirements in a Jira ticket using a multi-agent LangGraph workflow.

Features

• Multi-Agent Evaluation: Utilizes chained specialized agents (Parser, Analyzer, Reasoner, TestGen, Synthesizer).
• MCP Native: Integrates with Model Context Protocol to securely surface Jira and GitHub Data to agents.
• Explainable AI: Maps PR diff snippets back to specific Jira Acceptance Criteria.
• Confidence Scoring: Heuristic-based scoring on pass/fail ratio of requirements.
• Premium Dashboard: Neon-dark aesthetic built with Next.js and Tailwind.

Project Structure

• backend/: FastAPI API server and LangGraph agents.
• frontend/: Next.js Web Dashboard.
• docs/: Architecture diagrams and deeper documentation.

Running Locally

1. Prerequisites

• Python 3.10+
• Node.js 18+
• OpenAI API Key

2. Backend Setup

1. Change to the backend directory: cd backend
2. Create virtualenv: python -m venv venv
3. Activate: .\venv\Scripts\activate (Windows) or source venv/bin/activate (Mac/Linux)
4. Install requirements: pip install -r requirements.txt (or if not present, install manually listed in script)
5. Create .env file from snippet in backend/ and insert OPENAI_API_KEY. (Jira/GitHub keys are optional for mock testing).
6. Run server:

   uvicorn main:app --reload

3. Frontend Setup

1. Change to frontend directory: cd frontend
2. Install dependencies: npm install
3. Check and install explicit dependencies: npm i lucide-react
4. Run server:

   npm run dev

4. Usage

Open http://localhost:3000. Enter a valid PR URL and Jira Ticket (or use the mock default). Click Evaluate!

Note: For Hackathon demo purposes, without Jira tokens, querying for SPEC-123 will trigger a robust mocked requirement response flow.