Show HN: Deta Surf – an open source and local-first AI notebook (github.com)

Deta has open-sourced Surf, an “AI notebook” desktop app for macOS, Windows and Linux that brings local files and web content directly into a single, searchable thinking workspace. Built with Svelte, TypeScript and Rust, Surf stores data locally in an open Surf Flat File System (SFFS) and supports PDFs, web pages, YouTube, tweets, images, tables and more. Key features include split views and tabs, integrated citations (deep‑linked to page sections, video timestamps or PDF pages), web-search-triggered notes, YouTube and PDF Q&A, drag‑in media parsing, interactive no-code applets called Surflets, and offline use. Users can bring their own API keys, add cloud models, or run local language models; the codebase is available on GitHub and released under Apache 2.0 (with an MPL-2.0 exception for one patched dependency). Deta also offers a separately managed build that integrates with their servers under different terms. For the AI/ML community, Surf matters because it prioritizes local-first data, open formats and provenance—enabling reproducible, privacy-conscious research and rapid prototyping without constant window‑switching or manual scraping. Its support for local LLMs and clear citation links helps with model grounding and source tracking, while Surflets lower the barrier to creating interactive experiments and visualizations. The open-source release invites contribution, auditing and integration with research toolchains, while the license and hosted variant clarify tradeoffs between fully local deployment and Deta’s managed features.