Coming 2026 · macOS Native

Your footage cockpit.

From memory card to LTO tape. One app for ingest, analysis, assembly, and archive.


Post-production is 10 tools
duct-taped together.

/01

Hours lost to manual file wrangling

Copying cards, renaming clips, building folder structures by hand. Every project, the same ritual. Time you should spend cutting.

/02

No overview of your footage

12 cards, 400 clips, 3 cameras. What's sharp? What's usable? You won't know until you scrub through everything in the NLE.

/03

Archive anxiety

LTO tapes without catalogs. Projects from 2019 somewhere on a shelf. Finding that one shot means mounting tapes and hoping.


Six stages. One app.

RushDeck handles everything before and after the creative edit. We call it Sandwich Automation: automate the bread, keep the meat manual.

1

Project Setup

Configurable folder templates with LTO-aware tape splitting. Briefing dialog captures client, project type, format, target length and tone — before the first clip lands. Define your structure once, reuse it on every project.

2

Media Ingest

Verified copying with MD5 checksums and streaming pipeline — 148 MB/s sustained, beating tools like Canister. Health checks on every file via ffprobe. Auto-routing by camera, codec, or card. Resume after crash. Edit proxies and analysis proxies generated in parallel.

3

Analysis + Selects

On-device scoring via Apple Vision Framework: sharpness, stability, exposure, motion. Whisper transcription on your GPU at 15x realtime. Interview vs. B-roll classification at 88% recall, 94% precision — no cloud needed. Optional: VLM tagging and semantic clustering via your own API key or local LLM.

4

Timeline Assembly

AI-generated structure proposals based on your briefing and analysis. Select timelines and draft timelines with in-app preview. Export to Premiere XML, FCPXML, or DaVinci Resolve via OpenTimelineIO. Your rough cut is waiting when you open the NLE.

5

Review

Frame.io integration with smart comment batching — configurable timeout or manual "done reviewing" trigger. Local playback for internal checks. Approve/reject workflow that feeds back into your timeline. Push drafts to clients, pull feedback into your edit.

6

LTO Archive

32x8MB streaming ring buffer for sustained writes. Write-through catalog, tape-aware splitting for large projects. Resume after power loss or crash. Portable SQLite database that lives on the tape itself. Import foreign tapes from Canister or other tools. Search your archive without mounting a single reel.


Features that don't exist yet.

B-Roll Quality Scoring

Every shot scored by sharpness, stability, exposure, and motion at 6x realtime — entirely on-device via Apple Vision Framework. Sort 400 clips by quality in seconds. Optional VLM tagging adds semantic descriptions and CLIP-based similarity clustering.

Interview Auto-Selects

Whisper transcription at 15x realtime on Apple Silicon. Identifies questions, groups repeated takes semantically, picks the best version by clarity and completeness. Multicam pairing across cameras via transcript matching. Start your edit with the best take, not the first take.

Semantic Archive Search

Search your LTO tapes the way you think. "Sunset drone shot over river" finds the clip — even if you never tagged it.

Portable Catalog on Tape

A SQLite database travels with every LTO tape. Hand a tape to another editor, and they get the full catalog — no external database needed.


Native. Private. Yours.

macOS native. No Electron. No browser tab.

Apple Silicon optimized Metal GPU acceleration Works fully offline Swift / SwiftUI POSIX I/O — 290 MB footprint Crash recovery built in

Three AI tiers: run local with Ollama, bring your own API key, or subscribe to RushDeck AI. Same features, your choice. On-device analysis (Whisper, CV scoring, classification) works without internet. We never see your footage, your prompts, or your invoices.


Buy it. Own it. Choose your AI.

RushDeck

$199
one-time purchase · + $69/year for updates (optional)

Full pipeline: Project Setup, Ingest, Analysis, Timeline Assembly, Review, LTO Archive. Includes local AI (Whisper transcription, CV scoring, classification) — no internet, no API key, no subscription needed.

Cloud AI — three ways, your choice:

Local LLM
Free
  • Run Ollama on your own Mac
  • VLM tagging, semantic search, structure proposals
  • Your data never leaves your machine
  • Requires M1+ with 16 GB+ RAM
Bring Your Own API Key
Free
  • Use OpenAI, Anthropic, or Google directly
  • You pay the provider, not us
  • ~$5-15 per typical project (300 clips)
  • Full control over model and cost
RushDeck AI
$4.99 — $14.99/mo
  • Starter: up to 500 AI analyses/month
  • Pro: unlimited
  • No API keys, no setup — just works
  • We handle the routing and the billing

Every option uses the same AI features. There is no feature gating — only a choice in how you connect.


Built by editors, for editors.

RushDeck comes from Flavored Media — 13 years of video production, thousands of projects delivered, too many hours lost to file management. We're building the tool we always wanted. Not a startup chasing a trend, but a production company solving its own problems and sharing the result.