LandPPT is an AI-powered presentation generation platform that automatically converts documents and topics into professional PowerPoint slides using large language models.
The tool addresses the time-consuming process of creating presentations by automating the entire workflow from outline generation through final export. Users input a topic or upload documents in formats including PDF, Word, Markdown, Excel, or PowerPoint. The platform generates an editable outline, then produces HTML slides with support for multiple templates and AI-adaptive layouts. It integrates web research capabilities through Tavily and SearXNG to supplement content with current information, handles image sourcing from local libraries, public image services, or AI generation, and produces speaker notes with text-to-speech narration. The system supports multiple AI providers including OpenAI, Anthropic Claude, Google Gemini, Azure OpenAI, and OpenAI-compatible interfaces like DeepSeek and Ollama, with per-role model routing to control costs.
Adoption suits teams and individuals who need to generate presentations at scale or frequently. The platform works well for content creators, educators, and business users who want to focus on content rather than design and formatting. It handles both one-off presentations from topics and batch processing of existing documents. The tool offers flexibility in deployment, running locally with SQLite and in-memory caching for testing, or scaling to production with Docker Compose orchestration including PostgreSQL, Valkey, and MinIO. Export options span PDF, HTML, PPTX, images, DOCX, and Markdown formats, with optional public sharing links that include audio narration and subtitles.
Development shows consistent activity with regular updates addressing both core functionality and operational concerns. The project maintains comprehensive documentation and actively manages issues. The codebase demonstrates attention to production requirements through support for multiple deployment architectures and database backends. The team has built out authentication options including local accounts, OAuth via GitHub and Linux Do, and email verification. Recent work includes refinements to document parsing quality, image service capabilities, and the async task execution system for PDF and video generation.
