← Back to Projects
leptons1618@anishgiri.dev:~/projects/localrag cat project.toml
LocalRAG
Chat with your own documents without anything leaving the machine — Streamlit in front, LangChain and ChromaDB behind, and a local Ollama model doing the thinking.
- category
- AI / ML
- year
- 2024
- status
- Active development
- stack
- 7 tools
- highlights
- 6
01Highlights
- Streamlit for the interface, so the entire application is Python — there is no second process to start.
- LangChain's Ollama and Chroma integrations handle the model and the retriever, rather than talking to either one directly.
- `pypdf` and `pdfplumber` are both in the requirements, which is the scar tissue from getting PDF text extraction right.
- Switching models is a config change, and the model update path is written down in `MODEL_CONFIG_UPDATE.md`.
- A `diagnostic.py` and a `CHANGELOG.md` — the empty-retrieval bug is documented rather than quietly patched around.
- A lighter Vue query page sits alongside the Streamlit app for when the full interface is more than you need.
02Built with
- 01Python
- 02Streamlit
- 03LangChain
- 04ChromaDB
- 05Ollama
- 06pypdf
- 07pdfplumber