Find My Thingy: A Local-First AI Memory Assistant Powered by Gemma 3 4B

Description and use

Find My Thingy lets users keep documents locally and ask natural‑language questions, returning answers with source citations, eliminating manual file searching.

Problem

Users struggle to locate specific information within their own documents, often spending time manually opening files and searching for keywords.

Outcome

The application enables local, privacy‑preserving retrieval and question answering over user documents, reducing manual search effort and providing source‑cited answers.

Workflow

• Add documents (PDF, TXT, Markdown)
• Extract text, chunk, and create embeddings
• Retrieve relevant passages for a query
• Generate answer with Gemma 3 4B via Ollama, citing sources

Evidence

Find My Thingy is a local-first personal memory assistant that helps you find information stored in your documents using natural-language questions.

Source

The workflow has four main steps: Add documents, Extract and index, Retrieve relevant information, Generate an answer.

Source

Gemma 3 4B, running through Ollama, generates an answer grounded in the retrieved information.

Source

The application retrieves relevant passages and uses a locally running AI model to generate an answer with source citations.

Source