Insight Vault is a local, privacy-focused library for building LLM-based applications. It allows you to store, search, and summarize text completely offline.
With Insight Vault, you can:
- Search your local knowledge base.
- Chat with your documents interactively.
- Summarize large sets of information into concise outputs.
All data stays on your machine, with no external API calls required.
For more details see the documentation.
- Local Inference — Uses LLAMA for local LLM inference.
- Local Embeddings — Embeddings are created using SentenceTransformers.
- Privacy-first RAG — Store and query documents locally with ChromaDB.
- Interactive CLI — Intuitive CLI interface for searching, managing, and summarizing.
The following dependencies are required to run Insight Vault:
- Ollama — For local LLM inference.
- ChromaDB — Local document storage and vector database.
- SentenceTransformers — Embeddings for better document search and query matching.
To install Insight Vault, you can use the following command:
pip install insightvaultInsight Vault can be used via a simple CLI interface or as a Python library.
If you want to use the chat or summarize commands, you need to make sure that Ollama is running first.
Adding Documents
insightvault manage add-file <path_to_document>This command will add a file from the specified path to the local document database. You can also directly add text. The manage command is also used to list all documents, and to delete all documents.
Searching Documents
insightvault search "Explain RAG pipelines"This will search your indexed documents for the query “Explain RAG pipelines” and return the most relevant results.
Chat
insightvault chat "Explain RAG pipelines"This uses RAG which means it takes in a natural language query and returns a response in natural language based on the most relevant documents you have indexed.
Summarizing Documents
insightvault summarize "Explain RAG pipelines"Summarizes the text you provide. The flag --file can be used to summarize a file.
Insightvault provides three apps as part of the library:
RAGApp— For RAG pipelines.SearchApp— For searching indexed documents.SummarizerApp— For summarizing text.
For example, to use the SummarizerApp, you can do the following:
from insightvault import SummarizerApp
app = SummarizerApp()
await app.summarize("This is a loooong test")See the API Documentation for more information.
If you want to contribute to Insight Vault or run it locally for development, follow these steps.
- Clone the Repository
git clone https://github.com/daved01/insightvault.git
cd insightvault- Install Development Dependencies
pip install -e ".[dev]"This installs all required libraries for development, testing, and quality checks.
- Run Tests
pytest testsThis runs all the unit and integration tests to ensure everything is working properly.
- Run Quality Checks
We use ruff, mypy, and pre-commit hooks to ensure high code quality.
# Type checking with mypy
mypy insightvault
# Linting and formatting with ruff
ruff check . --fix
ruff format .Pre-commit Hooks
To automatically check for secrets, format code, and run linters before every commit, set up pre-commit hooks as follows:
pre-commit installThis will install the hooks and run them automatically before each commit.
To publish a new version of Insight Vault to PyPI:
- Update the version in pyproject.toml.
- Build the package:
python -m build- Publish the package:
twine upload dist/*We welcome contributions of all kinds. Whether it’s bug fixes, new features, or improving the documentation. Please open an issue or submit a pull request.