Avatar Video With Animated SFX Intro

I want to add a task that will generate a short 1 minute video of a talking head avatar reading out a promotional script, preceded by some animated video with sound effects. The video will begin with a short animation related to the script, with sound effects, and then the main video will begin with the talking head. Here's how it will work... first user inputs information about the video they want, and hits go. Next, GPT4 writes a short 800-1200 character script. Next, we turn the script into a voice-over mp3. Then we browse from a collection of "talking head images" and choose the one that is most relevant to our script (I will give the agent a folder URL containing images for it to choose from). Now we pass the selected talking-head.jpg image, along with the mp3 voice, to the talking head task which makes our video. Now we need to make our animated intro video. To do this we should generate an animated video. Next we generate our sound effects. Finally, we take the animated video, sound effect file and talking head video and combine them into a timeline to make our final video.

First, pre-entered info...

final-output2

The final output of this automated workflow would be a short promotional video in MP4 format, approximately 1 minute in length. The video will consist of two main parts: 1. An animated intro (approx. 3-6 seconds) with relevant sound effects 2. A talking head avatar video (approx. 45-50 seconds) lip-syncing a promotional script These two parts will be combined into a single MP4 video file, stitched together in a timeline.

subagent-summary-amended-for-actual-tasks

A) SUBAGENT SUMMARY: Generate a talking head video of an AI avatar lip-syncing an input MP3 voice-over. B) FINAL TASK OUTPUT: An MP4 video file of an AI-generated talking head avatar lip-syncing the input MP3 voice-over. The video duration will match the length of the input MP3 file. C) SUBAGENT INPUT: - MP3 URL containing the voice-over audio (from Subagent 2) E) SUBAGENT TASK SUMMARY: The subagent will complete the following sequence of skills to generate the talking head video: [script-voiceover-mp3] > #219 - Generate Talking Head Video From MP3 > [talking-head-video] Detailed breakdown: 1. The subagent receives the MP3 URL from the previous subagent (Voice-over Creator) as its input. 2. Skill #219 is used to generate the talking head video. It takes the MP3 URL as input and generates an MP4 video of an AI avatar lip-syncing the audio. 3. The resulting MP4 video URL is the final output of this subagent, saved as [talking-head-video]. F) SILOS: This subagent does not require any silos as it consists of a single skill that directly generates the desired output from the given input.

subagentnumber

OVERVIEW OF SUBAGENTS: SUBAGENT 1: Script Generator - Skill Used: #178 - Rewrite Text Based On Instructions - Description: This skill uses a LLM (typically GPT4) to generate a short promotional script based on user input and instructions. - Input Needed: Structured user input data from the user interface - Output: A text file containing a 120-180 word generated promotional script - SUBAGENT FINAL OUTPUT: [generated-promo-script] SUBAGENT 2: Voice-over Creator - Skill Used: #187 - Turn Script Into Voice Over MP3 - Description: This skill accepts a text script and converts it to an MP3 voice-over file. - Input Needed: Text script from Subagent 1 - Output: MP3 File URL with both internal and external link - SUBAGENT FINAL OUTPUT: [script-voiceover-mp3] SUBAGENT 3: Talking Head Video Generator - Skill Used: #219 - Generate Talking Head Video From MP3 - Description: This skill will input an MP3 and generate a "talking head" video of an AI avatar lip-syncing the audio. - Input Needed: Voice-over MP3 URL from Subagent 2 - Output: URL of the talking head avatar MP4 video - SUBAGENT FINAL OUTPUT: [talking-head-video] SUBAGENT 4: Intro Animator - Skill Used: #194 - Make Short Animated Video From Text Prompt - Description: This skill will return a short 5 second AI-generated animation video based on a given prompt derived from the script. - Input Needed: The generated script from Subagent 1 - Output: MP4 URL of the animated intro video with both internal and external link - SUBAGENT FINAL OUTPUT: [animated-intro-video] SUBAGENT 5: Video Stitcher - Skill Used: #207 - Concatenate Videos (Combine 2-10 MP4s) - Description: This skill stitches and concatenates multiple MP4s into one single video file. - Input Needed: Animated intro MP4 URL from Subagent 4, and talking head avatar MP4 URL from Subagent 3 - Output: The final composed MP4 video URL with both internal and external link - SUBAGENT FINAL OUTPUT: [final-promo-video] SUBAGENT #6: Video Outputter - Skill Used: #215 - Summarise and Return All Final Output Files - Description: This skill returns the final video file along with the individual MP4s of the animated intro and avatar. - Input Needed: Animated intro MP4 URL from Subagent 4, and talking head avatar MP4 URL from Subagent 3 - Output: simple text summary of the links - SUBAGENT FINAL OUTPUT: [video-links] SUBAGENT SEQUENCE: 1. Script Generator 2. Voice-over Creator 3. Talking Head Video Generator 4. Intro Animator 5. Video Stitcher 6. Video Outputter

Then, agent response...

subagent1refined-new-json-core

G) TASKFLOW SUMMARY - ID: #219 - Title task: Generate Talking Head Video From MP3 - Description of the task: This task takes an MP3 voice-over as input and generates an MP4 video of an AI-generated talking head avatar lip-syncing the audio. The video duration matches the length of the input MP3 file. - Input(s) involved: [script-voiceover-mp3] - Output token name: [talking-head-video]

subagent1mermaid

<!DOCTYPE html> <html> <head> <title>Mermaid Diagram Example</title> <script src="https://cdnjs.cloudflare.com/ajax/libs/mermaid/9.4.3/mermaid.min.js"></script> </head> <body> <div class="mermaid"> graph TD A[MP3 URL] --> B{Generate Talking Head Video From MP3 #219} B --> C[talking-head-video] </div> <script> mermaid.initialize({ startOnLoad: true }); </script> </body> </html>

Mermaid Image - https://static.aiz.ac/1720296482-mermaid/mermaid1.png

prompts-required

Based on the provided taskflow summary, there is only one skill involved in this subagent, and it does not require an LLM prompt: ID #: 219 Task: Generate Talking Head Video From MP3 Output of the task/skill: An MP4 video file of an AI-generated talking head avatar lip-syncing the input MP3 voice-over. The video duration will match the length of the input MP3 file. Description: This task is a pre-built skill that directly generates the desired output (talking head video) from the given input (MP3 voice-over). It does not involve an LLM or require any specific prompt to be written.

subagent1-json-array

[ { "taskID": 219, "semanticTitleOfTask": "Generate Talking Head Video From MP3", "taskDescription": "This task takes an MP3 voice-over as input and generates an MP4 video of an AI-generated talking head avatar lip-syncing the audio. The video duration matches the length of the input MP3 file.", "inputDescription": "The MP3 URL containing the voice-over audio from the previous subagent (Voice-over Creator).", "inputRequired": ["script-voiceover-mp3"], "outputDescription": "An MP4 video file URL of an AI-generated talking head avatar lip-syncing the input MP3 voice-over. The video duration will match the length of the input MP3 file.", "outputName": "talking-head-video", "promptInstruction": "" } ]

subagent1-json-array-mermaid

<!DOCTYPE html> <html> <head> <title>Mermaid Diagram Example</title> <script src="https://cdnjs.cloudflare.com/ajax/libs/mermaid/9.4.3/mermaid.min.js"></script> </head> <body> <div class="mermaid"> graph TD A[script-voiceover-mp3] --> B{Generate Talking Head Video From MP3} B --> C[talking-head-video] </div> <script> mermaid.initialize({ startOnLoad: true }); </script> </body> </html>