Presenton

Presenton is an open-source AI presentation generator and API, positioned as a self-hosted alternative to tools like Gamma, Canva, Beautiful AI, and Decktopus. It lets users create presentations from a simple prompt, an uploaded document, or an existing PowerPoint design, then export fully editable PPTX or PDF files. There is no SaaS lock-in or forced subscription, giving teams complete control over their models and data.

Architecturally, Presenton can run as a fully self-hosted Docker package, a native Electron desktop app for Windows, macOS, and Linux, or through Presenton Cloud directly in the browser. It supports a wide range of model providers including OpenAI, Google Gemini, Vertex AI, Azure OpenAI, Amazon Bedrock, Anthropic, Ollama, LM Studio, and any OpenAI-compatible endpoint, so users bring their own API keys and only pay for what they use. A built-in MCP server also allows presentation generation over the Model Context Protocol.

Key features include:

• Custom Templates and Themes — build unlimited designs with HTML and Tailwind CSS
• Flexible Generation — create decks from prompts, documents, or existing PowerPoint files
• Bring Your Own Key — connect any supported LLM or image provider
• Versatile Image Generation — DALL-E 3, Gemini Flash, Pexels, or Pixabay
• Multi-User Workspaces — private workspaces with an admin panel
• API Deployment — host Presenton as a team-wide API service

The project ships built-in templates like Momentum, Dynamic, Executive, General, Modern, and Standard for varied use cases. It is licensed under Apache 2.0, supports GPU-accelerated Docker deployment for local models, and integrates presentation memory via Mem0 alongside document parsing through LiteParse.