Deta Surf is an open-source AI notebook app that integrates your files, web content, and thinking process into a unified workspace.
It’s designed for those moments of deep research and thought, cutting down on the tedious tasks of switching between windows, searching for files, and copy-pasting content into a separate editor.
Think of it as an alternative to Google’s Notebook LM that provides a central hub where all your research materials, be it PDFs, YouTube videos, or web articles, can be organized and analyzed with the help of AI.
Features
Multi-Media Library Storage
Save websites, YouTube videos, tweets, Reddit posts, PDFs, images, and local files in a private library on your computer. The tool creates local files, database entries, and vector embeddings for text content, all stored in transparent formats through the SFFS (Surf Flat File System) storage engine.
Smart Notes with @-Mentions
Generate notes by @-mentioning any tab, website, PDF, or resource in your library. The AI uses these mentions as context to produce rich notes with citations that link directly to specific pages in PDFs, timestamps in videos, or sections on webpages.
Web Search Integration
Toggle web search within notes to research topics without leaving your notebook. Results come back with citations to the web pages where information was found.
Offline Access
View saved items offline once they’re added to your library. Right-click any resource and select “View Offline Version” to access content without an internet connection.
Surflets (Code-Free Apps)
Ask the AI to generate interactive visualizations, charts, games, or tools directly in your notes. These mini applications can store data locally between uses and are saved as HTML files on your machine.
Split View & Sidebar Navigation
Work with multiple resources side by side. Click citations in notes to open sources in split view, or use the sidebar for quick access to notebooks and saved materials.
Flexible AI Model Support
Use built-in models with your own API keys (OpenAI, Anthropic), add custom models through OpenAI-compatible endpoints, or run local models through Ollama. The app includes model switchers for changing between options on the fly.
Automatic Content Recognition
The app identifies and handles specific content types like YouTube videos, tweets, PDFs with full text embedding, and images with automatic OCR.
Use Cases
Academic Literature Reviews: Open multiple research papers and recorded lectures, then ask Surf to compare methodologies, extract limitations, and generate comparison charts with all sources cited. Everything stays in an offline-capable notebook for continued work without internet access.
YouTube Research Projects: Save multiple videos on a topic to your library, then @-mention them in a note to ask comparative questions. Citations link directly to relevant timestamps, letting you jump to exact moments where information appears.
Cross-Format Documentation: Combine information from PDFs, web articles, Notion documents, Google Docs, and Slack messages in a single research project. The AI can pull insights from all formats without manual copying and pasting between applications.
Data Visualization Without Coding: Ask Surf to create interactive graphs, physics demonstrations, or data exploration tools using natural language. The generated Surflets can persist data locally, useful for todo trackers, games with saved scores, or ongoing analysis projects.
Multilingual Content Analysis: Process research materials in foreign languages and scientific diagrams, with AI assistance to summarize and explain complex content in accessible terms.
How to Use It
1. Download the latest stable release from the GitHub releases page for your operating system (MacOS, Windows, or Linux). Alternatively, download the managed version from deta.surf, which comes with integrated AI features but operates under different terms.
2. Click the blue “Surf” button from any tab to save the content to your library. For local files, click the “media” tab and select “Import Files.” The app automatically detects content types and creates appropriate local storage.
3. Organize related resources into Notebooks by right-clicking items and selecting “Add to Notebook.” You can customize notebook covers by right-clicking a notebook and using the scribble feature.
4. Start typing in a new tab and hit “ask Surf” to create an AI-generated note. From an existing tab, click the “ask” button in the top right to open split view with that tab added as context. Use @-mentions to reference specific resources, tabs, or other notes in your library.
5. Go to Preferences → AI to set up your preferred models. For built-in providers, enter your API keys from OpenAI or Anthropic. For custom models, click “Add Model” and provide the OpenAI-compatible API endpoint URL, model ID, and optional API key.
6. Toggle web search at the bottom of notes to let AI fetch information from the internet. Enable the Surflet tool to generate code-based visualizations or interactive elements within your notes.
7. Search your entire library from any new tab. Right-click any resource to manage it, view it in your file system, or access its offline version.
Shortcuts
| Shortcut | Action |
|---|---|
| General Navigation | |
⌘ + T | Create a new tab or go to the command menu. |
⌘ + E | Toggle the notes sidebar. |
⌘ + O | Change tab bar orientation (horizontal/vertical). |
⌘ + , | Open app settings. |
| Browsing | |
⌘ + R | Reload the current tab. |
⌘ + 1-8 | Switch to a specific tab. |
⌘ + W | Close the active tab. |
⌘ + Shift + T | Restore the most recently closed tab. |
⌘ + ← | Navigate backward in history. |
⌘ + → | Navigate forward in history. |
⌘ + Shift + C | Copy the URL of the active tab. |
⌘ + L | Focus and edit the active URL. |
| View | |
⌘ + + | Zoom in. |
⌘ + - | Zoom out. |
⌘ + = | Reset zoom to default. |
| Emergency | |
Shift + ⌥ + ⌘ + R | Restart the entire application. |
Pros
- Complete Data Ownership: All information stays on your device in open, transparent formats.
- Model Flexibility: Works with commercial APIs, custom hosted models, or local LLMs through Ollama. You’re not locked into a single AI provider or forced to send data to external servers if using local models.
- True Offline Capability: Continue working on notes and accessing your library without internet connection.
- Open Source Core: The main codebase is available on GitHub under Apache 2.0 license.
- Deep Citations: Generated notes include clickable references that take you to exact locations in source materials (specific PDF pages, video timestamps, webpage sections) rather than generic source links.
- Cross-Platform Support: Consistent experience across MacOS, Windows, and Linux systems.
Cons
- Setup Complexity for Custom Models: Configuring your own AI endpoints requires understanding technical details like context windows, vision support, and JSON output capabilities.
- Limited Built-In Models: You need to bring your own API keys for popular models like GPT-4 or Claude. There’s no free tier AI included with the open-source version.
- No Mobile Version Yet: The app is desktop-only.
Related Resources
- Deta Surf GitHub Repository: Official source code, documentation, and community contributions for the open-source project.
- Ollama: Recommended tool for running AI models locally on your machine, compatible with Surf’s custom model configuration.
- OpenRouter: Hosted AI model service that provides access to multiple models through a single API, works with Surf’s OpenAI-compatible endpoints.
- Hugging Face Inference Providers: Platform for accessing text generation models through routing services like Together AI, compatible with Surf’s custom model setup.
FAQs
Q: Does Deta Surf require an internet connection to work?
A: No, once you’ve added items to your library, you can view them offline by selecting “View Offline Version” from the resource menu. However, generating new AI notes, using web search, or accessing online content requires internet.
Q: Can I use Deta Surf without paying for AI services?
A: Yes, if you set up local models through Ollama or similar tools. The open-source version doesn’t include free AI credits, so you either need to use local models or bring your own API keys from providers like OpenAI or Anthropic.
Q: Where exactly does Surf store my data?
A: Everything lives on your local machine in the SFFS (Surf Flat File System). You can see the exact location by selecting “Show Surf Data in Finder/Explorer” from the app menu. Each item gets stored as a local file, database entry, and vector embedding. You can also right-click individual resources and select “View in Finder/Explorer” to see their specific files.
Q: How does Surf compare to Google’s NotebookLM?
A: Deta positions NotebookLM as a direct competitor, arguing that Surf’s tab context and local data storage make it better suited for students and researchers. The main differences are Surf’s browser integration, offline capability, and local-first architecture. NotebookLM requires uploading sources to Google’s servers, while Surf keeps everything on your device.
Q: What happens to my notebooks if I switch computers?
A: You’ll need to manually transfer your Surf data folder to the new machine, since everything stores locally. Cloud backup and multi-device sync are features Deta may include in a future premium subscription, but they’re not available in the current version. For now, backing up your Surf data folder is your responsibility.
