A prompt management application with profile and persona-based templating for AI interactions. Create reusable personas, prompt templates with variable substitution for various use cases.

• Supports Social Login - OAuth 2.0 flow
• Support Multi Profile for one Logged in User
• Complete Isolation of user profile, templates and prompts.

User can play a role and may want LLM to play another role. For example:

• I am a High School Student and want LLM to play role of Patient Teacher. Some other time same person may want LLM to play a role of Mock Tester or Test Question Generator or may be a Student Buddy. In each scenario, different prompts, appropriate to the role play will yield best results.

• Multi-Profile for Logged in User: One logged in user can have multiple profile and subsequent persona, template and prompts
• User Level Isolation: Complete isolation of data at user level.
• Persona Management: Define user and LLM roles for different contexts
• Template System: Create reusable prompt templates with {{variable}} placeholders
• Template Version Management: Maintain versions of templates and edit them separately. Track generated prompts
• Prompt Generation: Generate final prompts by substituting template variables
• Web Interface: Modern SvelteKit frontend for easy management
• REST API: Full CRUD operations for all entities
• Storage Options:
  • JSON Store: Simple file-based persistence (default)
  • SQLite Store: Database storage with automatic schema initialization
• Docker Support: Containerized deployment with ECR integration
• Production Ready: EC2 deployment with docker-compose

• Generic RAG Application: User adds their artifacts and later wants to query that corpus. However they want to use sophesticated prompts to customize response. Example: A High School student may want to generate Mock Tests for his preparation.
• Interviewee: User adds their artifacts and later wants to query that corpus. They want to simulate interview process using LLMs ability to play a role of an interviewer.
• Pair Programming: User may want to use LLM for pair programming by asking LLM to play that role.

• Go 1.22+
• Node.js 18+

1. Clone the repository:

git clone https://github.com/rahulguha/promptly.git
cd promptly

2. Build the Go backend:

go build -o promptly cmd/promptly/main.go

3. Start the API server:

./promptly serve

4. Set up the web interface (in a new terminal):

cd promptly-web
npm install
npm run dev

The API will be available at http://localhost:8080 and the web interface at http://localhost:5173.

# Start server on default port 8080 with JSON storage
./promptly serve

# Custom port and data path
./promptly serve --port 3000 --data ./custom-data/prompts.json

# Use SQLite storage
./promptly serve --storage sqlite --db data/promptly.db

# Custom port with SQLite
./promptly serve --port 3000 --storage sqlite --db ./custom-data/app.db

# Help
./promptly --help

1. Create Personas: Define user roles (e.g., "Developer") and corresponding LLM roles (e.g., "Code Reviewer")
2. Build Templates: Create prompt templates linked to personas with variable placeholders like {{project_name}}
3. Generate Prompts: Substitute variables with actual values to create final prompts

• GET/POST/PUT/DELETE /v1/personas - Manage user/LLM role definitions
• GET/POST/PUT/DELETE /v1/templates - Manage prompt templates
• GET/POST/PUT/DELETE /v1/prompts - Manage generated prompts
• POST /v1/generate-prompt - Generate prompts from templates
• GET /health - Health check

# Run tests
go test ./...

# Build
go build -o promptly cmd/promptly/main.go

cd promptly-web

# Development server
npm run dev

# Build for production
npm run build

# Preview build
npm run preview

1. Build the binary for Docker:

./docker/build-api-mac.sh

2. Build and run with Docker:

# Build the Docker image
cd docker && docker build -f dockerfile.promptly-api-local -t promptly-api-local .

# Run locally
docker run -p 8082:8082 -v "$(pwd)/data":/app/data promptly-api-local

1. Deploy using docker-compose:

# On your EC2 instance
sudo mkdir -p /opt/promptly/data
cd /opt/promptly

# Copy docker-compose.yml and configure
wget https://raw.githubusercontent.com/yourusername/promptly/main/docker-compose.yml

# Create production config
echo "port: 8082
data: data/prompts.json" > config.yaml

# Login to ECR (configure AWS CLI first)
aws ecr get-login-password --region us-east-1 | docker login --username AWS --password-stdin your-account.dkr.ecr.us-east-1.amazonaws.com

# Start the service
docker-compose up -d

The application looks for configuration in:

• Command line flags (--port, --data, --storage, --db)
• Config file (config.yaml or .promptly)
• Environment variables

JSON Storage (default):

• Data directory: data/ (creates prompts.json, prompt_template.json, persona.json)
• Use: --storage json --data path/to/data.json

SQLite Storage:

• Database file: data/promptly.db (auto-created with schema)
• Use: --storage sqlite --db path/to/database.db

• API Documentation - Complete REST API reference
• Examples - Usage examples and demo data

See CONTRIBUTING.md for guidelines on how to contribute to this project.

Licensed under the Apache License 2.0. See LICENSE for details.