hughminhphan

hughminhphan / vector-embedded-finder

Public

Local multimodal memory with semantic search. Gemini Embedding 2 + ChromaDB + Raycast extension.

11
1
100% credibility
Found Mar 20, 2026 at 11 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

A local tool for embedding and semantically searching images, audio, videos, PDFs, and text files using natural language, including a Mac menu bar extension for visual results.

How It Works

1
🕵️‍♂️ Discover the memory finder

You hear about a simple tool that helps you search your old photos, videos, and documents just by describing them in words like 'sunset beach'.

2
💻 Bring it to your computer

You download it to your Mac and set it up with a few easy steps, including linking a free smart service to understand pictures and sounds.

3
📁 Add your personal files

You show it folders of your photos, music clips, videos, PDFs, or notes, and it quietly remembers them all in one safe local spot.

4
🔍 Ask in plain English

You type a casual question like 'team dinner last year' and watch as the best matching images, videos, or pages pop right up.

5
🖥️ Quick search from menu bar

Using the handy Mac extension, you open a grid of thumbnails and find files visually in seconds with smart matching.

🎉 Memories at your fingertips

You effortlessly rediscover forgotten moments, open files instantly, and feel organized without digging through folders ever again.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 11 to 11 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is vector-embedded-finder?

This Python tool builds a local multimodal memory system, embedding images, audio, video, PDFs, and text files into a ChromaDB vector store using Gemini Embedding 2 for 768-dim vectors. You ingest files or directories via simple API calls like `ingest_directory("~/Photos")` or `ingest_file("team-dinner.jpg")`, then query with natural language like `search("sunset beach")` to get ranked matches across modalities—no cloud storage, all local. A Raycast extension adds Mac-native grid search with thumbnails and instant file opening.

Why is it gaining traction?

Cross-modal semantic search shines: text queries pull relevant photos or videos without metadata, powered by local multimodal embeddings in ChromaDB for fast cosine similarity. The Raycast integration delivers 400ms-debounced visual results, deduping via hashes, and it's dead simple to set up with a Gemini API key. Developers dig the privacy of local multimodal RAG without vendor lock-in.

Who should use this?

Mac devs with Raycast needing quick semantic search over personal media libraries, like photo hoarders querying "team dinner" to surface event clips. AI tinkerers building local multimodal AI chat prototypes or embedded RAG pipelines. Content creators indexing local multimodal assets for retrieval.

Verdict

Early alpha with 11 stars and 1.0% credibility score—solid README and MIT license, but expect rough edges like API rate limits on Gemini. Worth a spin for local multimodal experiments if you have Raycast; skip for production until more battle-tested.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.