PersonalRAGVault - Find the answer in your own files.
by•
Search PDFs, notes, code and saved exports on your machine. Inspect matching source text, or optionally generate answers with local Ollama. Free, MIT licensed, with a CLI, browser UI and an API. Try the synthetic sample without an API key.
Replies
Hi, I'm Sean, the builder behind NextEleven.
PersonalRAGVault brings local search to the files you already have: PDFs, notes, code and saved exports. You can inspect the matching text and source before deciding whether to ask a model for an answer.
It is free under the MIT license. The CLI, optional browser interface and local API support different ways of using the same retrieval workflow. Local Ollama is the default generation option. An external compatible provider is an explicit choice and receives retrieved context for generated answers.
Start with the included synthetic sample. It uses real embeddings and hybrid retrieval, requires no API key, and shows the fictional invoice amount alongside its source. Installation and the first model download need internet access.
This is a local Python application, not a managed hosted service. Retrieval quality on your own collection and generated-answer correctness need evaluation; the sample is not a benchmark. The matching v1.5.1 release now includes wheel and source downloads with checksums: https://github.com/seanebones-lang/personal-RAG/releases/tag/v1.5.1. Start with the pinned source tag or extracted source archive and run the included synthetic sample.
I'd value feedback on first use, source inspection, and searches that miss something useful. Please use synthetic examples in public bug reports so private documents stay private.
Built by NextEleven.