StoryToolkitAI
patreon
The Story Editor is here
🕑 Added 2023-12-19 19:45:13 +0000 UTCIn our last post, we mentioned that we're going to focus our attention on creating a system that (1) allows the selection of content from transcriptions and (2) lets AI auto-complete with its own suggestions based on existing content.
After weeks of tinkering and testing in our editing room, step 1 is (almost) complete!...
So, what's the Story Editor?
Under the form of a new window, the Story Editor allows you to select text either from transcripts or from search results (right click - Add to story) and basically have them in a screenplay form, while preserving the timecode data and source information for later export.
Once you've made your selections, you can simply export the story, either as a timeline (or sequence) in EDL or FCP7XML, which can be imported into Resolve, Premiere Pro, Avid MC, etc. The Story Editor also supports exporting to Fountain screenplay format in case you want to import it into compatible screenwriting software, or for printing after additional conversion (to PDF for eg.)
We tried to make the export-import experience as easy as possible, but EDL and FCP7XML files do come with limitations and additional conforming might be required - please let us know how things work on your end!
For detailed instructions, see this. A video how-to will follow soon!
Who is this for?
We think that the Story Editor feature might come handy not just for editors, but also for writers, directors or producers who simply want to have a rough cut based on interviews, or a few topic-based select reels that they can export and send to their editors.
What's next?
We still need to fix a few things here and there, and then launch the standalone version of the tool with this new features.
After that, we're going to try to figure out exactly what else we need to enable the AI based auto-complete feature (step 2 mentioned before). Given the recent advancements with large language models (especially OpenAI GPT 4 Turbo, GPT Assistants and GPT-Vision), we're going to make an up to date assessment into what could be a good AI-based workflow, and which of the available tools can be implemented fast. By the way, if you're already using the Assistant feature of the tool, or just text AIs in general and feel that your workflow makes sense, we'd love to hear from you!
A huge request from many users is the ability to recognize speakers during the transcription process. There are solutions, but unfortunately none were good enough to consider, except for the ones that require some sort of paid service. We're constantly looking into this and will definitely implement it as soon as we find the right fit!
We still need to make a few things clearer and easier to understand, for eg. did you know that you can search project.json files for Resolve markers? (we're sure that there are a lot of things we haven't really explained well yet...)
How to update (git only for now)
If you're on the git version, a simple git pull will do the trick. Standalone users, we're almost there – just need to fix some things before!
Heads up on bugs
Since we're only able to test this on a handful of machines and software, we rely, as always, on the community to let us know if something isn't right. If you notice any timecode drifts, missing clips, or anything else that's not documented, shouldn't be there or just annoying, drop us a line before deleting the tool or trashing your computer!
This is already too long!!
Thanks again for the patience and your support! We're going to try to keep updates more frequent (and shorter) in the weeks to come!
Oh, and by the way, Happy Holidays!! 🥳🎉🪩
