Deploy and Host Presentation AI on Railway

About Hosting Presentation AI

Presentation AI is an open-source, AI-powered alternative to Gamma for creating professional slides. It features an outline-first workflow, real-time slide generation, customizable built-in themes, rich text editing via Plate Editor, and presentation recording. It is designed for creators, educators, and teams looking to automate slide creation efficiently.

Common Use Cases

AI Assistants & Content Creators: Generate complete, structured presentation decks automatically from simple text prompts.
Internal Tools & Business Dashboards: Create reusable branded slide templates and present directly within organizations.
Developer Tools & Education: Customize an open-source presentation engine using custom themes, local LLM integrations, and rich text modules.

Dependencies for Presentation AI Hosting

Dependency	Purpose
PostgreSQL	Database storage for application data, themes, users, and slide content
OpenAI API	Cloud text model provider for generating slide outlines and slide content
Together AI API	Cloud model provider for generating presentation images
UploadThing Token	Service for handling media file uploads
Unsplash Access Key	Provider for sourcing stock presentation images
Tavily API	Web search integration for retrieving real-time external context

Deployment Dependencies

GitHub Repository

Implementation Details

Docker

Railway automatically builds and deploys the application directly from the source code repository.

Public Networking

Proxy Type	Target Port
HTTP Proxy	3000

Environment Variables

Variable	Required	Description
`DATABASE_URL`	Yes	PostgreSQL connection string format (`postgresql://user:password@host:port/dbname`).
`NEXTAUTH_SECRET`	Yes	Secret key used for authenticating session tokens.
`NEXTAUTH_URL`	Yes	Canonical base URL of the deployed application.
`OPENAI_API_KEY`	No	API key for OpenAI text generation models. Optional if using local models.
`TOGETHER_AI_API_KEY`	No	API key for Together AI image generation models.
`UPLOADTHING_TOKEN`	No	Access token for UploadThing file upload service.
`UNSPLASH_ACCESS_KEY`	No	API key for Unsplash stock image integration.
`TAVILY_API_KEY`	No	API key for Tavily web search integration.

DATABASE_URL=postgresql://postgres:postgres@localhost:5432/presentation_ai
NEXTAUTH_SECRET=development_secret_key_123456789_presentation_ai
NEXTAUTH_URL=http://localhost:8080
OPENAI_API_KEY=
TOGETHER_AI_API_KEY=
UPLOADTHING_TOKEN=
UNSPLASH_ACCESS_KEY=
TAVILY_API_KEY=

API keys for external services can be obtained directly from their respective developer dashboards:

OpenAI API key from the OpenAI Platform dashboard.
Together AI API key from the Together AI console.
UploadThing token from the UploadThing developer portal.
Unsplash access key from the Unsplash Developer dashboard.
Tavily API key from the Tavily AI portal.

Persistent Storage

Persistent storage is optional. The application relies on external integrations like UploadThing for handling media assets and cloud storage.

Database

A PostgreSQL database service is required.

Add a PostgreSQL database service in your Railway project canvas.
Connect the database service to the application using Railway reference variables (e.g., ${{Postgres.DATABASE_URL}}).
Ensure DATABASE_URL is configured in the environment variables table.

Build & Start

No custom build or start commands are required.

Accessing the Application

Open Railway Service Settings.
Under Networking, click Generate Domain to assign a public access URL.
Open the generated access URL in your browser.
Sign in to the application to set up your account and begin generating presentations.

Why Deploy Presentation AI on Railway?

Railway provides a modern, developer-friendly platform for hosting web applications and databases. With features like instant deployments from GitHub, automatic HTTPS, built-in environment variable management, and seamless scaling, Railway minimizes infrastructure overhead so you can focus on building your application.