Music Stems from YouTube Links

I want a simple agent which downloads 3 YouTube videos (subagent1) and then extracts instrumentals and vocals from each (subagent 2). Finally, it returns a summary of all the files - the youtube url, mp3 download url of the song, and the vocals and instrumentals, in a single summary response (subagent3).

First, pre-entered info...

final-output2

The final output would be a single, straightforward file (for example, a text file) that includes, for each of the three YouTube videos, the original video link, a link or file reference to the downloaded MP3, and two additional links or file references for the separated vocals and instrumentals.

subagents-extracted2

OVERVIEW OF SUBAGENTS • Subagent 1: "YouTubeToMP3" – Downloads each YouTube video's audio as an MP3. • Subagent 2: "AudioSeparator" – Separates the MP3 into vocal and instrumental stems. • Subagent 3: "SummaryCompiler" – Compiles the final text summary from all collected data. SUBAGENT 1: "YouTubeToMP3" • Skill Used: #193 - Download Entire YouTube Video As MP3 x3 • Skill Used: #223 (Powerful LLM Prompt-to-Text Response) – to unify multiple MP3 outputs if needed • Input: Each YouTube URL - {variable1}, {variable2}, {variable3} • Output: MP3 file URL(s) for each downloaded video • Process: 1. Take each provided YouTube link. 2. Call Skill #193 to download the audio as MP3. 3. Since multiple MP3 files are generated (one for each video), call Skill #223 to gather the multiple MP3 outputs into a single textual output reference. 4. Return a unified list or text output of all MP3 URLs, mapped back to the original YouTube link(s). SUBAGENT REQUIRED INPUT: the 3 YouTube URLs - {variable1}, {variable2}, {variable3} SUBAGENT FINAL OUTPUT: [downloaded-mp3-urls] SUBAGENT 2: "AudioSeparator" • Skill Used: #197 - Extract Instrumental & Vocal From MP3 x3 • Skill Used: #223 (Powerful LLM Prompt-to-Text Response) – to unify multiple vocal/instrumental outputs if needed • Input: One (or multiple) MP3 URL(s) returned from Subagent 1 • Output: Two MP3 URLs (vocals, instrumentals) for each of the 3 input tracks • Process: 1. Take the MP3 URL(s) from Subagent 1. 2. Call Skill #197 to separate each MP3 into vocal and instrumental stems. 3. Since multiple sets of outputs are generated (e.g., for three videos), optionally call Skill #223 to gather the vocal/instrumental URLs into a single textual output reference per track (or combined). 4. Repeat for each of the 3 tracks. 5. Return the collated references (vocals and instrumentals URLs), mapped to each original MP3. SUBAGENT REQUIRED INPUT: [downloaded-mp3-urls] SUBAGENT FINAL OUTPUT: [vocals-instrumentals-urls] SUBAGENT 3: "SummaryCompiler" • Skill Used: #185 - Write Text From Inputted Text • Input: – Original YouTube URLs – Respective MP3 URLs – Vocals and Instrumentals URLs – A prompt outlining the desired format • Output: – A text summary ("summary.txt") listing all URLs and file references • Process: 1. Gather all URLs from Subagents 1 and 2, plus the original YouTube links. 2. Create the final summary text ("summary.txt") in a concise format, listing the original video link, downloaded MP3 link, vocals link, and instrumentals link for each track. 3. Save or return this summary text as the final single-file deliverable. SUBAGENT REQUIRED INPUT: {variable1}, {variable2}, {variable3}, [downloaded-mp3-urls], [vocals-instrumentals-urls] SUBAGENT FINAL OUTPUT: [final-summary-text]

subagentnumber

[subagentnumber]

Then, agent response...

subagent1refined-new-json-core

I'll add the G) TASKFLOW SUMMARY section to SUBAGENT 1, extracting and detailing each task/skill from the workflow: A) through F) [Previous sections remain exactly as provided] G) TASKFLOW SUMMARY 1. First YouTube Video Download - ID: #193 - Title task: Download First YouTube Video as MP3 - Description of the task: Downloads the audio from the first YouTube URL and converts it to MP3 format, storing it with a unique URL reference - Input(s) involved: {variable1} - Output token name: [mp3-url-1] 2. Second YouTube Video Download - ID: #193 - Title task: Download Second YouTube Video as MP3 - Description of the task: Downloads the audio from the second YouTube URL and converts it to MP3 format, storing it with a unique URL reference - Input(s) involved: {variable2} - Output token name: [mp3-url-2] 3. Third YouTube Video Download - ID: #193 - Title task: Download Third YouTube Video as MP3 - Description of the task: Downloads the audio from the third YouTube URL and converts it to MP3 format, storing it with a unique URL reference - Input(s) involved: {variable3} - Output token name: [mp3-url-3] 4. Consolidation of MP3 URLs - ID: #223 - Title task: Consolidate Multiple MP3 URLs into Structured Output - Description of the task: Takes all three MP3 URLs and their corresponding YouTube links, creates a structured mapping between them, and returns a unified text output containing all references - Input(s) involved: [mp3-url-1], [mp3-url-2], [mp3-url-3], {variable1}, {variable2}, {variable3} - Output token name: [downloaded-mp3-urls] This taskflow represents all skills used in the subagent, maintaining the original skill IDs while providing detailed information about each task's inputs, outputs, and purpose within the workflow.

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[Input: 3 YouTube URLs] --> B[First URL Process] A --> C[Second URL Process] A --> D[Third URL Process] B --> E[Task #193: Download First Video] C --> F[Task #193: Download Second Video] D --> G[Task #193: Download Third Video] E --> H[MP3 URL #1] F --> I[MP3 URL #2] G --> J[MP3 URL #3] H --> K[Task #223: Consolidate URLs] I --> K J --> K K --> L[Final Output: downloaded-mp3-urls] style A fill:#f9f,stroke:#333 style L fill:#9ff,stroke:#333 </div> <script> mermaid.initialize({ startOnLoad: true }); </script> </body> </html>

Mermaid Image - https://static.aiz.ac/1738363998-mermaid/mermaid1.png

prompts-required

[prompts-required]

subagent1-json-array

[ { "taskID": 193, "semanticTitleOfTask": "Download First YouTube Video as MP3", "taskDescription": "This task downloads the audio from the first YouTube URL and automatically converts it to MP3 format. The resulting audio file is stored with a unique URL reference for future use.", "inputDescription": "Requires the first YouTube video link (variable1). The skill retrieves the video from YouTube's servers, extracts the audio stream, and generates an MP3 file.", "inputRequired": ["variable1"], "outputDescription": "Returns an MP3 URL pointing to the newly created MP3 file, which can be referenced by subsequent tasks.", "outputName": "mp3-url-1", "promptInstruction": "no instruction" }, { "taskID": 193, "semanticTitleOfTask": "Download Second YouTube Video as MP3", "taskDescription": "This task downloads the audio from the second YouTube URL and automatically converts it to MP3 format. The resulting audio file is stored with a unique URL reference for future use.", "inputDescription": "Requires the second YouTube video link (variable2). The skill retrieves the video from YouTube's servers, extracts the audio stream, and generates an MP3 file.", "inputRequired": ["variable2"], "outputDescription": "Returns an MP3 URL pointing to the newly created MP3 file, which can be referenced by subsequent tasks.", "outputName": "mp3-url-2", "promptInstruction": "no instruction" }, { "taskID": 193, "semanticTitleOfTask": "Download Third YouTube Video as MP3", "taskDescription": "This task downloads the audio from the third YouTube URL and automatically converts it to MP3 format. The resulting audio file is stored with a unique URL reference for future use.", "inputDescription": "Requires the third YouTube video link (variable3). The skill retrieves the video from YouTube's servers, extracts the audio stream, and generates an MP3 file.", "inputRequired": ["variable3"], "outputDescription": "Returns an MP3 URL pointing to the newly created MP3 file, which can be referenced by subsequent tasks.", "outputName": "mp3-url-3", "promptInstruction": "no instruction" }, { "taskID": 223, "semanticTitleOfTask": "Consolidate All MP3 URLs into a Structured Output", "taskDescription": "This task gathers all three MP3 URL outputs, ties them to their corresponding YouTube links, and builds a unified mapping that associates each video with its MP3 download URL.", "inputDescription": "Requires the three MP3 download URLs (mp3-url-1, mp3-url-2, mp3-url-3) and the original YouTube links (variable1, variable2, variable3). The task then merges them into a combined structure.", "inputRequired": ["mp3-url-1", "mp3-url-2", "mp3-url-3", "variable1", "variable2", "variable3"], "outputDescription": "Returns a formatted text output mapping each original YouTube URL to the corresponding MP3 download link. This structured reference is used for clarity and documentation.", "outputName": "downloaded-mp3-urls", "promptInstruction": "Create a structured mapping of YouTube videos and their corresponding MP3 downloads. For each entry, include: 1) The original YouTube URL 2) The corresponding MP3 download URL. Format the output as follows:\n\nVideo 1:\n - YouTube: [URL]\n - MP3: [URL]\n\nVideo 2:\n - YouTube: [URL]\n - MP3: [URL]\n\nVideo 3:\n - YouTube: [URL]\n - MP3: [URL]\n\nEnsure all URLs are properly formatted and accessible. Maintain consistent spacing and formatting throughout, with each video's information clearly separated." } ]

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 A1[Input: 3 YouTube URLs] --> B1[Download First Video] A1 --> B2[Download Second Video] A1 --> B3[Download Third Video] B1 --> C1[MP3 URL 1] B2 --> C2[MP3 URL 2] B3 --> C3[MP3 URL 3] C1 --> D[Consolidate URLs] C2 --> D C3 --> D A1 --> D D --> E[Final Output: Structured URL Mapping] style A1 fill:#e6f3ff style B1 fill:#f9f9f9 style B2 fill:#f9f9f9 style B3 fill:#f9f9f9 style C1 fill:#f0f0f0 style C2 fill:#f0f0f0 style C3 fill:#f0f0f0 style D fill:#f9f9f9 style E fill:#e6f3ff </div> <script> mermaid.initialize({ startOnLoad: true }); </script> </body> </html>