The AI Story Collaborator
Turns hours of dialogue-driven footage into rough cuts in your style and sends them straight to
A Better Way to Work

Trained on You

My Style trains on your own portfolio, the scripts, transcripts, and finished pieces that show how you tell a story, so the rough cuts follow your structure, your pacing, how you open, and what you let breathe. Without a style loaded it works like a competent story producer. With one, the first pass reads like yours. Agencies and studios keep a profile per brand or show, so every rough cut matches the right voice.
Ask for the piece in plain language (“build me a 14-minute cut about leaving home”) and a selects sequence lands in
Everything On-device

Every part of the pipeline, transcription, speaker identification, and the language model that does the analysis and style matching, runs on-device. Footage never leaves the machine, nothing requires an account, and there are no per-minute meters or hour caps, which matters for NDA'd client work and for anyone cutting daily volume.
After first launch,
What’s new in the latest version?
- Chat is much smarter. It reads the whole interview however long it is, answers in plain words with playable clip cards, offers follow-up questions, knows when not to give clips, and remembers what the editor is making.
- Correct the transcript in place. Double-click any word to fix it, insert a word the engine missed, split a paragraph, or give a line to a different speaker. Cmd+Z undoes each step. Media timing never changes, so every clip keeps its timecode.
- Collections work like a project. Check two or more interviews on the Projects page and Create Collection; every tab (Chat, Story Brief, Story Builder, Clips, Export) works across all of them at once.
- A connector for Claude (MCP). Per project, off by default. One click adds it to Claude Desktop; Claude Code, Cursor and VS Code connect too. A connected assistant can read the switched-on project's transcript and selects, search it, and add its own labeled selects. Media files never leave the Mac. Support note: the access panel states plainly that transcript text goes to Anthropic while it is on.
- Exports are named for the project and numbered ("Project – Selects 1"), every exported clip carries its words in the note, Story Builder sequences export from the Export tab, and quote sheets use the press format by default.
- Clips get generated titles, a first line and Show text everywhere, plus a word-level Trim sheet; speakers are named from the transcript after speaker identification, and hand-set speakers survive a re-run.
- Import multicam clips straight from the
Final Cut Pro browser. Select the clips, choose the File > Export XML... command, drop the resulting file in. No scratch timeline needed. An export with several multicams becomes one project per clip. - Projects and exports now live in the Doza Assist folder inside your Documents, where they can be backed up and/or synched via iCloud or other cloud storage provider. Existing projects migrate automatically on first launch.
- Send to FCP now supports the Creator Studio as well as Trial versions of
Final Cut Pro - Miscellaneous export accuracy and stability fixes (clip selections export exactly what's checked, quit/relaunch hardening).
How do I install
Doza Assist?
Doza Assist is available via FxFactory, our app store for visual effects, audio plugins and apps:
Once FxFactory is on your system, click the button below to download and install Doza Assist:
Version History
Doza Assist
Version 1.1.0MPEG-2 camera files (Sony XDCAM MXF, MPEG-TS) now report their frame rate correctly. Retranscription no longer strands the analysis. The player no longer covers the text in a compressed window.
Doza Assist
Version 1.0.46Import multicam clips via
Doza Assist
Version 1.0.45Doza Assist
Version 1.0.43Requires macOS 15 or later, and a Mac equipped with Apple silicon.
Please note that this product will not work unless FxFactory is installed on the system.