What Shall We Build Next?

Describe

Describe your task

Refine

Refine the plan

SubAgents

Review all agents

Deploy

Deploy your agent

A) SUBAGENT SUMMARY  
Generates a concise (100–300 word) voice-over script based on the user’s prompt or topic.

B) FINAL TASK OUTPUT  
A single text script (in plain text format) that is 100–300 words in length.

C) SUBAGENT INPUT  
• [variable1]: The user’s prompt or topic (for example, “An introduction to holiday sales at a department store”).

E) SUBAGENT TASK SUMMARY  
Below is the detailed flow of how this subagent (“ScriptScribe”) completes its mission, including specific skill references:

1) (Optional) Create Outline (Silo 1)  
   • Action (Skill #185 - Write Text From Input):  
     ─ Input: [variable1] plus any instructions like “Please outline the key points to include in the final voice-over script.”  
     ─ Output: A brief textual outline of key talking points.  
   (Note: This step is optional but can help structure the script.)

2) Generate Draft Script (Silo 2)  
   • Action (Skill #171 - Write Voice Over Script Based On Instructions):  
     ─ Input: Combine [variable1] (the user prompt) and/or the outline from Step 1 (if created), specifying “Please write a 100–300 word voice-over script.”  
     ─ Output: A text script (initial draft) in the 100–300 word range.

3) (Optional) Verify Script Length + Refine (Silo 3)  
   a) Check Word Count  
      • Action (Skill #223 - Powerful LLM Prompt-to-Text Response):  
        ─ Input: “Count the words in the following script and verify if it is within 100–300 words” plus the text script from Step 2.  
        ─ Output: A short text response with the estimated word count.  
   b) Rewrite if Needed  
      • If the word count is not within the 100–300 range:  
        – Action (Skill #190 - Write or Rewrite Text Based on Instructions):  
          ▪ Input: “Please revise this script to make it fall within 100–300 words” plus the original draft script.  
          ▪ Output: A newly shortened or expanded script that meets the word requirement.

4) Subagent Final Output  
   The final script text (100–300 words) is returned from the above steps.

F) SILOS  
• Silo 1: (Optional) Outline Creation  
  – Summarize or outline the main points of the user’s input.  
• Silo 2: Draft Script Generation  
  – Use Skill #171 to create the initial 100–300 word script.  
• Silo 3: (Optional) Script Verification and Revision  
  – Check final script length (via #223), and if necessary, adjust word count with #190.

––––––––––––––––––––––––––––––––––––––––––––––––––  
This completes the Subagent #1 (“ScriptScribe”) workflow to ensure a concise voice-over script is generated from the user’s prompt.

SubAgent #1 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
This subagent (“VocalBuilder”) takes a short text script (≈100–300 words) and transforms it into an MP3 voice-over file.

B) FINAL TASK OUTPUT  
• A single MP3 audio file URL, containing the spoken version of the input text script.

C) SUBAGENT INPUT  
• Required: The text script that will be converted to speech.  
• Optional: Any additional instructions about voice style or tone (if the skill supports it).

E) SUBAGENT TASK SUMMARY  
Below is the detailed step-by-step workflow that VocalBuilder will follow:

1) Validate/Pre-check Script (Optional):  
   • (No additional skill call needed unless special constraints or rewrites are required.)  
   • Ensure the script is within the acceptable length (under ~4000 characters).

2) Convert Script to MP3 (Essential):  
   • Use Skill #170 (“Turn Script Into Voice Over MP3”).  
   • Input: [script-text] (plus any voice style or accent preferences, if supported)  
   • Output: An MP3 file URL.

3) Return the MP3 File  
   • Subagent returns the voice-over as the final output: [voiceover-mp3].

F) SILOS  
• SILO 1: (Optional) Script Pre-processing  
   - Purpose: If needed, double-check or lightly edit the text.  
   - (Uses no extra skill unless a rewrite or content check is required, e.g., #190 for rewriting.)

• SILO 2: Voice-over Generation  
   - Purpose: The core of this subagent—convert the script to an MP3.  
   - Skill #170: “Turn Script Into Voice Over MP3”  
   - INPUT: The validated text script (+ optional style/tone instructions)  
   - OUTPUT: [voiceover-mp3]

The subagent’s final output is the MP3 file URL, ready to be used in subsequent video-assembly steps.

SubAgent #2 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
"AvatarVision" is responsible for creating a single AI-generated avatar image (PNG) in 16:9 aspect ratio based on the user’s description (e.g., "woman wearing a Christmas hat in a department store").

B) FINAL TASK OUTPUT  
A single PNG image URL ([avatar-image]) that reflects the user’s avatar description and is in a 16:9 aspect ratio.

C) SUBAGENT INPUT  
• The user's avatar description (e.g., "woman wearing a Christmas hat in a department store")

E) SUBAGENT TASK SUMMARY  
1) Rewrite User Prompt For Image Generation (#190 - Write or rewrite text based on instructions)  
   - INPUT: The user’s avatar description  
   - ACTION: Rewrite the user’s description to explicitly include instructions that the image should be generated in 16:9 aspect ratio and any additional style/context needed (for example, “Create a 16:9 aspect ratio AI avatar image of a woman wearing a Christmas hat in a department store”)  
   - OUTPUT: A revised text prompt (e.g., “Generate a highly detailed, 16:9 aspect ratio image: woman wearing a Christmas hat in a department store…”)

2) Create Image With 16:9 Aspect Ratio (#222 - Make Image)  
   - INPUT: The revised avatar prompt (from step 1)  
   - ACTION: Generate one AI avatar image in PNG format, using the specified style and 16:9 aspect ratio  
   - OUTPUT: PNG image URL (this is the final [avatar-image])

F) SILOS  
• SILO 1: Prompt Preparation (Step 1)  
• SILO 2: Image Generation (Step 2)

Once step 2 completes, the subagent returns the PNG image URL as [avatar-image], satisfying the final output requirement for this subagent.

SubAgent #3 - Diagram

Expand Flow

A) SUBAGENT SUMMARY  
This subagent (“VideoSync”) takes the MP3 voiceover file and the user’s generated avatar image, produces a synchronized transcription of the audio, and then uses those pieces to render a lip-synced video of the avatar speaking the script.

B) FINAL TASK OUTPUT  
A single MP4 file URL (16:9 aspect ratio) showing the AI-generated avatar “talking” with synchronized lip movement to the provided MP3 voiceover.

C) SUBAGENT INPUT  
1) [voiceover-mp3]: The URL of the MP3 file containing the narrated script.  
2) [avatar-image]: The URL of the AI-generated avatar image (generated at 16:9).

E) SUBAGENT TASK SUMMARY  
Step 1 → (#198 - Get Transcription Of MP3)  
• INPUT: [voiceover-mp3] (the MP3 URL)  
• PROCESS: Transcribe the audio into text (with timings).  
• OUTPUT: [transcription-text] (timed transcript).

Step 2 → (#168 - Generate Talking Head Video From MP3 & transcription)  
• INPUT:  
   • [voiceover-mp3] (the same MP3 URL)  
   • [transcription-text] (from Step 1)  
   • [avatar-image] (the AI-generated image)  
• PROCESS: Generate a lip-synced talking head video of the avatar speaking the transcribed text.  
• OUTPUT: [talking-head-video] (MP4 video URL).

F) SILOS  
• SILO 1 (Audio to Transcription): Use #198 to produce a transcription from the MP3 voiceover.  
• SILO 2 (Transcription + Avatar → MP4): Pass the MP3, the transcription, and the avatar image into #168 to produce the final lip-synced MP4 video.

4 Template & Links

Expand Flow

A) SUBAGENT SUMMARY  
This subagent will take the final talking-head MP4 and add burned-in subtitles (closed captions) to the video, producing a second MP4 with on-screen text synchronized to the speaker’s narration.

B) FINAL TASK OUTPUT  
A single MP4 video file that is identical to the original talking-head MP4 but now includes burned-in subtitles (closed captions) appearing in sync with the spoken audio.

C) SUBAGENT INPUT  
1. The final talking-head MP4 video URL (from Subagent 4).

D) SUBAGENT TASK SUMMARY  
Below is the step-by-step flow this subagent follows to generate the final MP4 with burned-in subtitles:

1. (INPUT) → Skill #207 - Get Transcription From MP4 Video URL  
   • Inputs:  
     - MP4 video URL (the completed talking-head video from Subagent 4).  
   • Output:  
     - A time-coded text transcription of the voiceover (with approximate 3–8-second segments).

2. Step 1 Output (transcription) + (INPUT) → Skill #199 - Add Images & Videos On Top Of Existing MP4  
   • Purpose: Use the transcription timings to place matching text overlays (subtitles) onto the original MP4.  
   • Inputs:  
     - Original MP4 video URL (the same from Subagent 4).  
     - The time-coded transcription from Skill #207.  
     - Instructions for formatting the on-screen subtitle text (position, font size, color, etc.).  
   • Output:  
     - MP4 URL of the newly rendered video with burned-in subtitles.

3. (OUTPUT) The final MP4 (with subtitles)

E) SILOS  
Although short, this subagent can be considered as two silos or “sub-subagents”:

• Silo 1: Transcription  
   - Uses Skill #207 to extract time-coded text from the final MP4.

• Silo 2: Overlay Subtitles  
   - Uses Skill #199 to overlay the text (from Silo 1) onto the original MP4, resulting in a video with subtitles.

That concludes the complete flow for this subagent (sometimes referred to as “Subagent 5”), whose job is to add on-screen captions to the final talking-head video.

5 Template & Links

Expand Flow

A) SUBAGENT SUMMARY:  
There is currently no defined Subagent #6 in the workflow. Because “no agent found for subagent 6” is given as the only information, this indicates that Subagent #6 does not exist or has not yet been created. Therefore, there is nothing concrete to implement or decompose for Subagent #6 at this time.

B) FINAL TASK OUTPUT:  
No output is produced because there is no Subagent #6.

C) SUBAGENT INPUT:  
No input is needed, as there is no subagent to run.

E) SUBAGENT TASK SUMMARY:  
No tasks are performed, as Subagent #6 is not defined in the workflow.

F) SILOS:  
There are no silos or subtasks to define because Subagent #6 does not exist in the sequence of steps.

────────────────────────────────────────────────────
Since Subagent #6 is not part of the established workflow, the best we can do at this time is confirm that there is no further action to specify. If, in future, a Subagent #6 is defined (for example, to add background music, captions, or some other functionality to the final video), it will then require its own input → skill(s) → output flow.