Hemingway

Hemingway analyzes prompts, identifies issues, and suggests clearer rewrites across multiple language models. It includes a Next.js frontend and a FastAPI backend.

Overview

The application breaks prompt review into a few focused analysis steps:

• extract instructions
• detect general issues
• find contradictions
• check structure and formatting

It can then generate a revised version of the prompt and suggest improvements.

Features

Analysis

• Extracts explicit instructions from a prompt
• Flags clarity, structure, and effectiveness issues
• Detects contradictory requirements
• Checks formatting and overall prompt structure

Optimization

• Rewrites prompts based on detected issues
• Provides concrete improvement suggestions
• Supports multiple model providers, including OpenAI, Anthropic, Google, and Moonshot

Interface

• Displays results in separate analysis views
• Shows analysis progress while processing
• Exports output as plain text, Markdown, or Python
• Compares original and rewritten prompts side by side

Frontend

• Responsive layout for desktop and mobile
• Light and dark theme support
• Animated transitions
• Accessible UI components built with Radix UI

Architecture

Frontend

frontend/
├── app/                    # Next.js App Router pages
├── components/             # React components
│   ├── ui/                 # shadcn/ui components
│   ├── input-view.tsx      # Prompt input interface
│   ├── results-view.tsx    # Analysis results display
│   └── export-dialog.tsx   # Export functionality
├── lib/                    # Utilities and types
│   ├── api-client.ts       # Backend API integration
│   ├── types.ts            # TypeScript definitions
│   └── export-utils.ts     # Export helpers
└── hooks/                  # Custom React hooks

Backend

api/
├── src/
│   ├── app.py              # FastAPI application
│   ├── custom_agents.py    # AI agent definitions
│   ├── models/
│   │   ├── instruction.py
│   │   ├── issue.py
│   │   ├── improvement.py
│   │   └── prompt.py
│   └── utils.py
└── requirements.txt

Tech Stack

Frontend

• Next.js 15
• React 19
• TypeScript
• Tailwind CSS
• Radix UI
• shadcn/ui
• GSAP
• Lucide React

Backend

• FastAPI
• Python 3.8+
• OpenAI Agents
• litellm
• Pydantic

Installation

Requirements

• Node.js 18+
• pnpm
• Python 3.8+
• API keys for any model providers you want to use

Frontend

cd frontend
pnpm install
pnpm dev

Frontend runs at http://localhost:3000.

Backend

cd api
pip install -r requirements.txt
cp .env.example .env
python src/app.py

API runs at http://localhost:4000.

Environment Variables

Create api/.env:

OPENAI_API_KEY=your_openai_api_key
ANTHROPIC_API_KEY=your_anthropic_api_key
GOOGLE_API_KEY=your_google_api_key
MOONSHOT_API_KEY=your_moonshot_api_key
FRONTEND_URL=<http://localhost:3000>

Usage

UI workflow

1. Enter a prompt.
2. Select a model.
3. Run analysis.
4. Review extracted instructions, issues, and suggested changes.
5. Export the result if needed.

API

curl -X POST "http://localhost:4000/api/analyze" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o",
    "prompt": "Your prompt text here"
  }'

Example response:

{
  "changes": true,
  "optimized_prompt": "Optimized version of your prompt...",
  "extracted_instructions": {
    "instructions": [
      {
        "instruction_title": "Main Task",
        "extracted_instruction": "Specific instruction text..."
      }
    ]
  },
  "general_issues": {
    "has_issues": true,
    "issues": [
      {
        "issue": "Issue description",
        "priority": "high",
        "snippet": "Problematic text",
        "explanation": "Why this is an issue",
        "suggestion": "How to fix it"
      }
    ]
  },
  "improvements": {
    "improvements": [
      {
        "description": "Improvement description"
      }
    ]
  }
}

Development

Frontend

cd frontend
pnpm dev
pnpm build
pnpm lint
pnpm start

Backend

cd api
python src/app.py

Supported Models

• GPT-4.1
• GPT-4o
• o3 Mini
• Gemini 2.5 Pro
• Gemini 2.5 Flash
• Claude 4 Sonnet
• Kimi K2

Rate Limiting

The API limits requests by model:

• 10 requests per 10 minutes
• returns HTTP 429 when the limit is exceeded

Security and Privacy

• API keys are provided through environment variables
• CORS limits frontend access to approved origins
• prompts are processed at request time and are not stored

License

MIT