MUZIM Vibe Search finds video moments from text descriptions
MUZIM presents Vibe Search as locating a scene at a specific timestamp from a text description rather than returning only a matching video. The tool also searches and organizes files across a user’s existing library.
TL;DR
- Vibe Search is presented as a way to retrieve a specific video timestamp from a remembered scene, as hasantoxr’s Vibe Search post and hasantoxr’s timestamp demo describe.
- MUZIM is in public beta, according to hasantoxr’s public-beta post, and its official product page lists desktop downloads for macOS and Windows.
- Smart Stacking groups duplicate, near-duplicate, and burst-shot photos, hasantoxr’s Smart Stacking post says.
- An online AI Agent can turn selected library material into a social thread, captions, a report, or a 30-second highlight concept, according to hasantoxr’s agent example, which notes that some jobs use Credits.
MUZIM’s feature list says each video receives an auto-generated title, key moments, and transcript highlights. Its privacy explanation says core file processing stays local, while AI Chat sends locally generated structured summaries rather than original files to outside models.
Vibe Search
The useful shift is from finding a clip to finding the part of the clip that contains the remembered event.
MUZIM says Vibe Search matches plain-language queries against text, speech, faces, scenes, and sound inside photos, video, audio, and documents. Its video search description positions titles, key moments, and transcript highlights as the metadata behind that retrieval.
Smart Stacking and Collections
Smart Stacking addresses photo review by grouping duplicate, highly similar, and burst-shot images. The Collections feature groups material by people, places, topics, and moments, and says one file can sit in several collections without creating disk duplicates.
Storage can span the computer, external drives, and supported cloud services, according to hasantoxr’s storage post. MUZIM’s import details name Google Drive, Dropbox, and iCloud, and say imports do not replace the existing setup.
Desktop app and credits
The core app is free, while MUZIM’s credit table puts advanced AI-agent work on OpenSoul credits: 1,000 for $20, 6,000 for $100, or 15,000 for $200. The same table estimates a deep-intelligence search at roughly 0.05 credits.
Its desktop requirements list macOS 14+ on M1+ hardware with 16GB of memory, or Windows 10+ with 24GB, plus 10GB of free storage on either platform.
MUZIM MCP
Alongside the online agent in hasantoxr’s agent example, MUZIM describes a separate MCP context layer. Its MCP section says it prepares relevant context locally before connecting tools including ChatGPT, Claude Code, and Codex, while the tools receive selected context rather than the original files.
MUZIM also claims that sending only the relevant context reduces unnecessary token use.