Custom Avatar Video Generator

I want to build an AI agent that generates avatar talking head videos. User will enter two inputs /variables to begin - a rough description of what the video should be about, and second - the type of avatar image I want (eg woman wearing christmas hat in department store). Then the agent will make the script, voice over, image and then video. Here's how I am thinking the subagents will work - first it will write a script for me (approx. 100-300 words). Secondly, it will turn that script into a voice over audio. Next, it will generate an image of my AI avatar (which should be topical and related to the theme of my video). Finally it will take this custom ai avatar image and the voice over file and generate the talking head video in question, and return the mp4 as the final output. In this way I can create custom AI avatar talking head videos on any topic, and with the avatar image themed to the topic.

First, pre-entered info...

final-output2

The final output is a single MP4 video featuring the AI-generated avatar "talking" through the script (voiceover synced to the avatar's mouth movements).

subagents-extracted2

OVERVIEW OF SUBAGENTS • Subagent 1: Script Generator • Subagent 2: Voiceover Generator • Subagent 3: Avatar Image Generator • Subagent 4: Video Assembly & Lip Sync Agent SUBAGENT 1: SCRIPT GENERATOR Name: ScriptGenAgent Final Output: A 100–300 word text script. Skills Used: • #171 - Write Voice Over Script Based On Instructions Sequence of Actions: 1) Receive the user's topic and desired style. 2) Use skill #171 to produce a concise script (100–300 words). 3) Return the script. SUBAGENT REQUIRED INPUT: {variable1} SUBAGENT FINAL OUTPUT: [script-text] SUBAGENT 2: VOICEOVER GENERATOR Name: VoiceOverAgent Final Output: An MP3 file (voiceover). Skills Used: • #170 - Turn Script Into Voice Over MP3 Sequence of Actions: 1) Receive the script. 2) Use skill #170 to synthesize and return an MP3 file. SUBAGENT REQUIRED INPUT: [script-text] SUBAGENT FINAL OUTPUT: [voiceover-mp3] SUBAGENT 3: AVATAR IMAGE GENERATOR Name: AvatarImageGen Final Output: A PNG avatar image (1024×1024). Skills Used: • #182 - Create Dalle Image (Transparent Square) Sequence of Actions: 1) Receive the user's avatar description. 2) Use skill #182 to generate one PNG image. 3) Return the image file URL. SUBAGENT REQUIRED INPUT: {variable2} SUBAGENT FINAL OUTPUT: [avatar-image] SUBAGENT 4: VIDEO ASSEMBLY & LIP SYNC AGENT Name: TalkingHeadAssembler Final Output: The finished MP4 video. Skills Used: • #198 - Get Transcription Of MP3 (With Timings) • #168 - Generate Talking Head Video From MP3 & transcription Sequence of Actions: 1) Take the MP3 from Subagent 2. 2) Use skill #198 to obtain a timed transcription of the voiceover. 3) Pass the MP3 and transcription, plus the avatar image, to skill #168 to generate a lip-synced MP4. 4) Return the final video. SUBAGENT REQUIRED INPUT: [voiceover-mp3], [avatar-image] SUBAGENT FINAL OUTPUT: [talking-head-mp4]

subagentnumber

[subagentnumber]

Then, agent response...

subagent1refined-new-json-core

Here's the updated subagent summary for SUBAGENT 1, including the new G) TASKFLOW SUMMARY section: SUBAGENT 1: SCRIPT GENERATOR A) SUBAGENT SUMMARY This subagent (ScriptGenAgent) generates a concise 100–300 word voice-over script based on the user's topic and style instructions. B) FINAL TASK OUTPUT [script-text] ⇒ A succinct text script (around 100–300 words). C) SUBAGENT INPUT {variable1} ⇒ The user's topic and desired style instructions in text form. D) SUBAGENT TASK SUMMARY 1) Receive {variable1} (the user's input describing the topic and style). 2) Apply Skill #171 ("Write Voice Over Script Based On Instructions") to produce a concise voice-over script (100–300 words). 3) Return the completed script as [script-text]. E) SILOS No separate silos are needed. This subagent's single objective is direct: generate the final script from the user's prompt. G) TASKFLOW SUMMARY Skill/Task #171: - ID: #171 - Title task: Write Voice Over Script Based On Instructions - Description of the task: Generates a concise, engaging voice-over script (100-300 words) based on user-provided topic and style preferences, optimized for spoken delivery and natural flow - Input(s) involved: {variable1} - Output token name: [script-text] [Note: This subagent uses only one primary skill (#171), which directly transforms the user input into the desired script output. The simplicity of this subagent's workflow means there's only one task to document in the taskflow summary.]

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[User Input: Topic and Style] --> B[ScriptGenAgent Start] B --> C[Process Input Instructions] C --> D[Apply Skill #171: Write Voice Over Script] D --> E[Generate 100-300 Word Script] E --> F[Quality Check: Verify Length and Style] F --> G[Return Final Script Text] G --> H[ScriptGenAgent End] style A fill:#f9f,stroke:#333,stroke-width:2px style H fill:#bbf,stroke:#333,stroke-width:2px </div> <script> mermaid.initialize({ startOnLoad: true }); </script> </body> </html>

Mermaid Image - https://static.aiz.ac/1747070804-mermaid/mermaid1.png

prompts-required

[prompts-required]

subagent1-json-array

[ { "taskID": 171, "semanticTitleOfTask": "Generate a concise 100–300 word voice-over script", "taskDescription": "This task uses skill #171 to read the user's topic and style instructions, then it writes a concise 100–300 word script suitable for voice-over delivery with a natural, conversational tone.", "inputDescription": "This task requires the user's topic and style instructions in text form, which guide the content, tone, and flow of the final script.", "inputRequired": ["variable1"], "outputDescription": "A 100–300 word script that addresses the user's topic and specified style, structured for voice-over with clear paragraphs and a natural flow.", "outputName": "script-text", "promptInstruction": "You are an expert script writer for voice-over videos. Your task is to create a concise, engaging script (100-300 words) based on: {variable1}. Use a natural tone, a clear introduction, body, and conclusion, short sentences, and subtle transitions. Aim for a rhythm that sounds natural aloud. End with a conclusive statement." } ]

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[Start] --> B[Receive User Input] B --> C[User Topic & Style variable1] C --> D[Task #171: Script Generation] D --> E[Apply Voice-Over Script Writing Skill] E --> F[Generate 100-300 Word Script] F --> G[Natural Flow & Tone Check] G --> H[Format Script for Delivery] H --> I[Return script-text Output] I --> J[End] style A fill:#f9f,stroke:#333,stroke-width:2px style J fill:#f9f,stroke:#333,stroke-width:2px style D fill:#bbf,stroke:#333,stroke-width:2px </div> <script> mermaid.initialize({ startOnLoad: true }); </script> </body> </html>