Quire

Quire is an open-source answer engine that lets you ask questions about your own documents in plain language and receive answers where every claim is linked to its source. Point it at a folder of documents, ask a question, and get grounded responses with citations, or an honest "not in your corpus" when information isn't available.

Quire is an answer engine you run yourself. You point it at a folder — contracts, invoices, a decade of internal wiki exports — and it builds a local index. You ask it a question in plain language. It answers, and every claim in the answer links back to the exact line it came from. If it cannot find support for an answer, it refuses and says so, rather than producing fluent text with nothing underneath.
The tool runs entirely offline on your local machine using local language models—no API keys, accounts, or cloud connectivity required. It supports multiple model providers including Ollama, llama.cpp, vLLM, OpenAI, and Anthropic, with flexible storage options via SQLite and pgvector. Quire indexes documents into a searchable database and provides instant, source-linked answers that you can verify.

Designed for teams, researchers, and individuals who need reliable, verifiable answers from their own document collections without relying on external services or cloud infrastructure.