Show HN: Swiftgum (Open Source) – Turn Data into LLM-Ready Markdown

Hey HN, We’re the co-founders of Swiftgum (swiftgum.com), an open-source platform that ingests and normalizes documents from virtually any platform, ready for your favorite LLM. Swiftgum lets you connect various data sources like Google Drive and Notion, transform the data into markdown, and load it into your vector database of choice, such as Postgres, Supabase, Milvus, or Weaviate. Why We Built It Building a basic RAG system is easy—just embed your documents and query them with an LLM. But for enterprise or multi-user scenarios, you quickly run into challenges: Data sources are messy. Files come in different formats, and converting them into a structured, AI-ready format is tedious. Users want to add their own documents to a shared knowledge base, creating potential conflicts or duplication. Permissions explode in complexity, especially when each user decides which docs to share (or not). Maintaining a consistent pipeline for ingestion, format conversion, and RBAC across multiple sources becomes a heavy lift. How Swiftgum Helps Open Source & Self-Hostable The codebase is available on GitHub (github.com/Swiftgum/swiftgum), so you can review, fork, or run Swiftgum on-prem for total control. No black boxes—you see exactly how data is ingested, transformed, and secured. Ingest & Normalize Any Document Swiftgum extracts, cleans, and converts documents from Drive, Notion, and other sources into LLM-ready Markdown. No more manual reformatting—just plug in your data and start querying. Per-User Sharing & RBAC Users decide what they want to share or keep private. Swiftgum enforces these settings in real time, ensuring only authorized persons can access specific documents. Flexible Data Export Instead of locking you into a specific database, Swiftgum lets you export data via webhooks, so you can send it to your preferred system. Whether it’s a vector database, a storage bucket, or a custom pipeline, you stay in control. Use Cases Enterprise AI Assistants & Knowledge Bases: Let every team member contribute documents while keeping private info private. SaaS Platforms: Provide a multi-tenant RAG service for your customers without custom-building complex permissioning logic. Compliance-Heavy Industries: Enforce fine-grained data visibility in finance, healthcare, and other regulated sectors—fully auditable and open-source. Why We’re Different Turns any data into LLM-ready Markdown. No need to clean, convert, or process documents manually. Truly multi-user. One centralized RAG pipeline with role-based access at every step. Open source & transparent. Unlike proprietary vector or document platforms, Swiftgum gives you full control. Try It or Contribute Get started with our hosted or on-prem options at swiftgum.mintlify.app/getting-started/quick-start. Check out the GitHub repo (github.com/Swiftgum/swiftgum) to spin up your own instance or contribute features. Our documentation (swiftgum.mintlify.app/introduction) explains how to connect data sources, configure user permissions, and integrate with your LLM environment. Feedback Wanted What data sources should we support next—SharePoint, Confluence, or something else? What tricky RBAC edge cases do you face that we can handle out-of-the-box? What deployment approach do you prefer—Docker, Kubernetes, or something else? We’d love your thoughts on making multi-user RAG simpler, more secure, and fully open-source. We’ll be here in the comments to answer questions—thanks for reading! — The Swiftgum Team