Product
Why FrameQuery Uses Hybrid Architecture (And Why Your Video Tool Should Too)
Your originals are never stored on our servers. Your search index lives on your disk. Why FrameQuery is a hybrid desktop app, cloud-processed and locally searchable, instead of another cloud platform.
Most video management platforms upload your footage to the cloud. FrameQuery is a desktop application. Your videos stay on your drives, your search index is a file on your disk, and search is free and works offline.
The cost of cloud-first video tools
A day on a RED or Blackmagic camera can produce hundreds of gigabytes. Uploading that takes hours or days and then costs storage fees every month.
Once uploaded, the footage lives on someone else's servers. If the company shuts down, raises prices or changes terms, the library goes with it.
Video professionals already own storage: NAS, RAID arrays, SAN. A second copy in the cloud is redundant.
How the hybrid architecture works
Processing (cloud). FrameQuery extracts sampled frames and audio on your device and sends only those to the indexing servers. That is true whether the source is a local drive, Google Drive or Dropbox. The extracted data is analysed (transcription, object detection, scene detection, scene captions), the results come back as an index, and the uploaded data is discarded. Face recognition and voice recognition run on your device and are never uploaded.
Search (local). The index is saved on your machine, at a few megabytes per hour of video. Searching it makes no network requests and costs nothing per query. It works with no internet connection.
You pay for processing, which is the GPU work on our servers. Search is free. If FrameQuery disappeared tomorrow, the index files would still work.
Rust and Tauri
We built the app on Tauri 2 with a Rust backend.
Rust. The backend decodes proprietary formats (R3D, BRAW), manages local databases, runs encoding pipelines and coordinates GPU processing. Rust gives C++ performance with memory safety, no garbage collector pause during a frame decode and no null pointer crash in a long batch.
Tauri. Tauri ships a native webview instead of a bundled Chromium. The installer is about 10 MB instead of 200 MB or more, memory use is a fraction of Electron, and CPU work runs at native speed instead of in a Node.js event loop.
Local database. Video records, source folders, processing status, settings and search indexes are a few files on disk. No server, no daemon, no container.
What offline-first means in practice
- Search the indexed library on a plane with no Wi-Fi
- Move index files to another machine and keep searching
- Back up FrameQuery data with ordinary file backup tools
- Restore from a backup without contacting any server
- Share an index with a collaborator who searches it on their own machine
The cloud is involved once, for indexing.
Encrypted backups
FrameQuery backs up to a local folder, Google Drive, Dropbox or a network folder or NAS, on demand or on a schedule. The library metadata is encrypted into one portable file. Restore it on a new machine and the library is back.
Trade-off
You are responsible for your own backups. If a drive dies and the index was not backed up, it is gone, and the videos can be re-processed to rebuild it. People who already manage petabytes of storage know how to back up files. A monthly cloud bill to search their own footage is the thing they should not need.
Download FrameQuery to try it.