Story AI processes your footage — transcripts with word-level timecodes, emotion scores, visual tags, speaker labels, markers. All of that lives in AVScript.
MetaFlow gets it into your NLE.
Not just Final Cut Pro. Every major professional editing suite has a native path: Final Cut Pro gets FCPXML with captions and Smart Collection keywords. Premiere Pro gets native XMEML — File → Import, no Translation Report, no XML bridge. DaVinci Resolve reads the same FCPXML. One analysis dataset. Three native export paths.
The NLE is your choice. MetaFlow adapts.
Three Paths, One Dataset
<caption> — every spoken word, frame-accurate, searchable in FCPThe same Story AI analysis — the same transcription job, the same emotion scores, the same visual tags — feeds all three. There’s no separate analysis step per NLE. You run the pipeline once, and MetaFlow exports to whichever NLE you’re working in today.
What MetaFlow Exports
Every source file that’s been through Story AI carries this data into the export:
| Data | Source | FCP | Premiere | DaVinci |
|---|---|---|---|---|
| Transcript | Transcribe | <caption> timecoded | V2 text generator | Subtitle track |
| Emotion peaks | Emotions AI | FCP Keyword | Bin description | — |
| Visual / topic tags | Visual AI | FCP Keyword | Bin description | — |
| Speaker labels | Transcribe | FCP Keyword | Bin description | — |
| Markers | AVScript | <marker> | — | — |
| Clip notes / overlay | AVScript editor | — | V2 text + bin description | — |
FCP gets the richest metadata because its XML format has native concepts for all of it — captions, keywords, markers. Premiere gets captions via V2 legacy text generators and all annotation text via <logginginfo><description> in the bin. DaVinci gets the timeline and captions; its import layer doesn’t surface FCP keyword collections, but the cut and the spoken word track land cleanly.
How to Export — Per NLE
What the FCP Keywords Look Like
When you import MetaFlow FCPXML into Final Cut Pro, every source clip gains keyword collections from the AI analysis:
Once they’re in FCP, Smart Collections do the heavy work. Build a Smart Collection for Emotion-Tension and instantly see every high-tension moment across your entire source library. Filter by Speaker-A to find every line from one interview subject. Every keyword from every clip in every project you’ve ever run through Story AI — organised and searchable in one place.
Premiere and DaVinci don’t have an equivalent to FCP’s keyword system, so those tags consolidate into the bin description text for Premiere, or aren’t surfaced in Resolve.
Source File Relinking
MetaFlow assets export with /RELINK/ placeholder paths — not server URLs, because all three NLEs relink to local files, not remote HTTPS addresses.
Before importing into any NLE: make sure your source files are available locally. If they’re stored in the AVScript cloud library, download them first. Each NLE shows its own version of a “missing files” dialog — navigate to each local file to relink.
After relinking, the NLE remembers the local paths. The metadata is permanently attached to your source files in that project.
How each NLE relinks:
- FCP — “Missing Files” dialog on first import. Navigate to each file.
- Premiere — “Link Media” dialog. Point at originals or proxies; Premiere matches by duration.
- DaVinci — “Relink Clips” dialog in the Media Pool. Navigate to local files.
FPS Mismatch During Import
If your NLE rejects a file with a frame rate mismatch, the frameDuration in the export doesn’t match what the NLE detected in the actual file.
AVScript handles the most common cases automatically:
- NTSC SD footage (480p at 29.97 fps) — auto-corrected. 480p is uniquely NTSC; if the file metadata incorrectly reports 25 fps, AVScript overrides to the correct NTSC rate.
- Library fps — at export time, AVScript uses the fps stored against each file in your cloud library, which is set when the file is first analysed.
If you still see a mismatch, use the FPS override in the export modal to set the exact rate manually before exporting again.
One Analysis. Any NLE. Forever.
This is the principle behind MetaFlow: AI data belongs to the file, not the project, and not the NLE.
When Story AI processes a video, the results — transcript, emotion scores, visual tags, speaker labels — are stored against the media file in your library, permanently. Not the project. Not the export destination. The file.
What this means in practice
Delete the project? MetaFlow data survives. Open a new project with the same file and all captions, keywords, and analysis are still there.Switch NLEs mid-production? No re-processing. The same file exports to FCP today and Premiere tomorrow at no extra cost.
Re-use a file in a new edit months later? MetaFlow exports immediately. You pay for work done, not work repeated.
The only way to lose the AI data is to delete the media file itself from your library.
Practical workflow: Run Story AI once when you first ingest a file. From that point on, every project that uses that file — in any NLE, in any edit, for any client — has MetaFlow available whenever you need it.
When MetaFlow Is Available
The MetaFlow checkbox in the FCPXML export modal is active when at least one bin file has transcript or AI data. For Premiere XML, the caption and description options are active when clip descriptions or overlay text exist.
No AI data yet? Run at minimum Transcription through Story AI → Pipeline. Transcript-only already produces captions in FCP and V2 text in Premiere. Emotion and visual tags appear as you run additional pipeline stages — just re-export when they’re ready.
MetaFlow doesn’t change how you edit. It makes everything AVScript learned about your footage available inside whatever tool you finish the cut in — permanently, across every project and every NLE that uses those files.
→ Run Story AI on your footage · → For FCP editors · → For Premiere editors · → Back to the manual