Rocketnotes is a web-based Markdown note taking app with LLM-powered text completion, chat and semantic search.
It utilizes a 100% Serverless RAG pipeline built with
langchain,
sentence-transformers,
faiss,
Ollama and OpenAI or Anthropic.
- Sign Up for free and use it in the Browser or as an Electron app.
- Run it 100% locally with Docker
- Host at AWS
- Check Contribution Guide how to setup a local development environment
- If you are interested in contributing, visit Contributing section for more details
- Good first issues provide a good starting point to become a contributor
- π Code syntax highlighting
- π Katex and Mermaid support
- π³ Hierarchical document tree with draggable nodes
- π Document sharing
- π Content search
- π¦ Semantic search
- βοΈ Copilot-like text completion
- π€ Chat with your documents
- Serverless RAG with faiss, OpenAI and/or Anthropic
- π¦ Local mode with Docker
- use Ollama and/or Sentence Transformers for 100% local RAG
- π₯ Zettelkasten with semantic archiving
- Use vector index to insert notes into highest semantic-ranked documents
- WIP: Neovim plugin rocketnotes.nvim
Β
- π€ Use the power of LLMs together with vector embeddings to chat with your notes or search them semantically.
- π€ Get Copilot-like text completion autosuggestions.
- βοΈ Save your daily note snippets into zettelkasten.
- π₯ Use vector index to insert notes into highest semantic-ranked documents with ease.
Β
- π Use the power and simplicity of Markdown for your personal notes.
- π» Create useful code snippets in your favourite programming language with syntax highlighting.
- π Share documents with external users.
Β
- π Search through all your documents by content.
- π Get an autosuggestion panel with all documents matching you search pattern - superfast!
Β
- π Save your note documents hierarchical with unlimited depth of subdocuments.
- ποΈ Structure your notes by simply drag and drop the desired document.
- π Pin favourite documents for fast top-level access.
A good way to do the first contribution are the good first issues. On some of the issues are already linked PRs with comments in the code what has to be done. These PRs can be picked up if not assigned to someone yet.
The most comfortable way to get started is to open the project in a ready-to-code Gitpod workspace with all packages & tools preinstalled and a running database with sample data.
Β
If you prefer to setup the project on your local machine see Contributing Guide and learn all required steps to run it locally in development mode.
Don't hesitate to open an issue for getting some feedback about a potential bug or if you desire a missing feature. It is appreciated to check over current issues and provide feedback to existing ones or even raise a PR which solves an issue. Any contribution is welcome!