Omni
Local-first multimodal file search on macOS
Details
- External ID
- 48419626
- Source
- HN
- Company
- —
- Product
- Omni
- Website domain
- hanxiao.io
- Launched
- June 5, 2026
- Cohort
- —
- Upvotes
- 6
- Upvotes percentile
- 0.31420765027322406
- Tags
- —
- Fetched at
- Sept. 7, 2026, 9:26 p.m.
- Updated at
- Sept. 7, 2026, 9:26 p.m.
Description
Finally made something I've always wanted, using the model we built.• SOTA omni embedding model, fully local, indexes text, PDF, image, audio, and video • Swift-native app UI + mlx-swift-transformer core. No Python. • Tested on M3 Pro 18G / M3 Ultra 512G / M4 Pro 48G. All work fine. • HTTP server exposes search to local agents like OpenClaw & Hermes − Indexing still feels slow even on the latest M3 Ultra, ranging from 10K tps to 300 tps depending on file type − Fans go crazy, high power draw while indexing − Search is near-instant. Multimodal relevance is sometimes arguable, but the idea is recall (the agentic LLM takes the results and refines for the final answer), so maybe that's fine
Enrichment
- Theme
- lightweight and on-device AI runtimes
- Vertical
- Horizontal
- Function
- Search & retrieval
- Audience
- B2C
- AI stance
- Not AI
- Project type
- Commercial product
- Normalized one-liner
- local file search for macos
- Manually corrected
- False
Could you build this?
Partial A developer can vibe-code the macOS Swift UI and local search orchestration, but developing a state-of-the-art unified multimodal embedding model from scratch requires deep ML research and training clusters.
What it would actually take: The production app uses Apple's MLX Swift runtime to execute matrix multiplications directly on Apple Silicon GPUs for local vector search. While the UI and SQLite/vector store indexing pipeline can be vibe-coded, the hard piece is the custom multimodal embedding model (spanning text, audio, image, and video in a single joint vector space), which requires large-scale contrastive pre-training across massive paired datasets using GPU compute clusters.
Discussion
2 comments analyzed.
Concerns raised: indexing performance is slow
Feature requests: run indexing as lower priority background task
Competitors
Other products that read as similar to this one — 52 launches clear the similarity bar, closest 8 shown.
Attention rank: #35 of 53 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 185 days after the earliest competitor.
- Open Chronicle · hn · 2026-04-22 · 5 upvotes · similarity 0.46
- Clipto · ph · 2026-05-31 · 522 upvotes · similarity 0.43
- Llama 3.2 3B and Keiro Research achieves 85% on SimpleQA · hn · 2026-03-07 · 6 upvotes · similarity 0.41
- Omni · hn · 2026-03-02 · 177 upvotes · similarity 0.41
- MetaBrain · hn · 2026-06-02 · 8 upvotes · similarity 0.39
- Omacosy · hn · 2026-08-20 · 57 upvotes · similarity 0.37
- Local text, image, video, music and 3D from one CLI, no Python · hn · 2026-07-30 · 16 upvotes · similarity 0.37
- Unified multimodal memory framework, without embeddings · hn · 2026-01-07 · 7 upvotes · similarity 0.36
Other launches for this product
- No other launches for this product.