The two days before, you shared with us the name of the funeralCoze worked to untangle and teach you to produce 100 short videos of explosive philosophy in a minuteAnd the psychological philosophy account, which is soared on the shivering。
NO, I FOUND ANOTHER AIHot VideosIn the case of this big guy, he's got 117 pieces, he's got 76,29,000 hits, and he's got 124,000 fans
Such videos, although streamlined, are philosophical, subject-oriented and resonant。
Matches the features of the short video, "light quantification, easy to communicate, high fit"。
I've been working on it all day, and I've got oneWorkflowIt's a very well-suited workstream, which, in addition to a single key, produces a short video of the blast in the Sensor language, it's also suitable for any recording, such as the Emotion track, the Enterprise track, the Healing track, the Psychopathic track, the Bitter Word, the awakening
Today, I'll teach you how to use itCozeCreate a working stream that can quickly generate all types of video。
It's a very practical article
Workstream Presentation
I'll split it up for you
Step one, node
I only have three set up here at the beginning
1. bottom_wenan: bottom text
for example:
2.mp3_url: uploading audio mp3 files
3.bj_wenan: background text
It's usually the subject, like "to meet."
So what we need to achieve at the end of the day is an intermediate deployment that you can enter directly into an option you want to share, or upload a self-defined text, and start automatically。
Step two, audio-video text
Step three, audio subtitle alignment
Align the script with the audio
step 4, str_to_list audio transfer list
String rotation list
step five, audio_timelines extract audio time lines
Step six, large model - background theme
It is built on a professional text subject to extract experts who can finely extract the core background theme from the user case。
System prompt words
User Hint
Step 7, Process background text data structure
Aggregation of outputs from multiple branches
Step eight. Words
This is a DeepSeek-V3.2 model
It's built with a [Chinese string cutter], not a writer or a content-generation model。
Only " pre-existing strings " are split, and nothing is allowed to be created, supplemented or rewritten。
System prompt words
User Hint
Step nine, time-axis alignment
Step 10, upload background pictures
In the end, cut the mix
The first selection of the subject from the big model, and then a mix of the drawings, the soundings, will converge in that plugin
And then there's a whole set of editing workflows that you're offering, and that's a theme that you're going to produce, and it's going to produce, and it's going to create, and it's going to sound, and it's going to sound, and it's going to just have to be directly exported, and you're going to see, you know, it's not complicated, and it's going to be built again。
WELL, THAT'S THE END OF THIS PERIOD'S AI。