What Shall We Build Next?

Describe

Describe your task

Refine

Refine the plan

SubAgents

Review all agents

Deploy

Deploy your agent

SubAgent #1 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
This subagent takes in a single MP3 file (the YouTube audio) and separates it into two stems—vocal and instrumental.

B) FINAL TASK OUTPUT  
Two MP3 URLs:  
• One MP3 URL for the vocal stem  
• One MP3 URL for the instrumental stem

C) SUBAGENT INPUT  
• [youtube-mp3-file]: the MP3 file URL generated by Subagent 1

E) SUBAGENT TASK SUMMARY  
1) [youtube-mp3-file] → (Skill #197: Extract Instrumental & Vocal From MP3) → [vocal-and-instrumental-stems]

• This step calls Skill #197 with the MP3 URL.  
   • Skill #197 outputs two separate MP3 URLs: one for the vocal track, one for the instrumental track.

2) Subagent produces final output ([vocal-and-instrumental-stems]) returning both stem URLs.

F) SILOS  
SILO 1: EXTRACT AND RETURN VOCAL + INSTRUMENTAL  
• Action: Invoke Skill #197 ("Extract Instrumental & Vocal From MP3") using the MP3 URL as input.  
• Output: Two MP3 URLs (instrumental, vocal).  
• Store/return these two URLs as the subagent’s final result.

SubAgent #2 - Diagram

Expand Diagram

A) SUBAGENT SUMMARY  
This subagent (BPMCALCULATOR) takes the MP3 file from Subagent 1 and returns the BPM (tempo) of that audio track.

B) FINAL TASK OUTPUT  
The final output is a single value: [bpm-value] (e.g. 128 BPM).

C) SUBAGENT INPUT  
• The only input is [youtube-mp3-file]: the MP3 URL from the Download Subagent.

E) SUBAGENT TASK SUMMARY  
Below is the detailed sequence of actions (skills) used to compute the BPM:

1) Take the [youtube-mp3-file] input.  
2) Use Skill #180 - “Extract Beatpoints & Tempo of MP3.”  
   INPUT to #180: [youtube-mp3-file]  
   OUTPUT from #180: A textual breakdown containing beatpoints, downbeats, and the identified tempo/BPM.  
3) Use Skill #223 - “Powerful LLM Prompt-to-Text Response.”  
   - This step parses the text output from Skill #180 to isolate the numeric BPM value.  
   INPUT to #223: The textual breakdown from #180 along with a request to “extract just the BPM number.”  
   OUTPUT from #223: [bpm-value] (a single numeric or short textual BPM value).

4) Final Subagent Output = [bpm-value]

F) SILOS

• Silo 1 (Beat Extraction):  
   - Skill #180 uses the MP3 URL → returns a textual breakdown with tempo info.

• Silo 2 (Parse BPM):  
   - Skill #223 takes that text → returns an isolated BPM value.

That completes the BPMCALCULATOR subagent’s flow.

SubAgent #3 - Diagram

Expand Flow

A) SUBAGENT SUMMARY  
This subagent receives the MP3 file URL, the vocal/instrumental stems, and the BPM value from earlier subagents, and uses a powerful LLM to compile a single cohesive text summary describing all three outputs.

B) FINAL TASK OUTPUT  
A single text response (string) that merges all relevant information, including:  
• The original MP3 URL ([youtube-mp3-file])  
• The two stem URLs (vocal and instrumental)  
• The BPM value of the original track

C) SUBAGENT INPUT  
• [youtube-mp3-file] – The MP3 file URL generated by the YouTube-to-MP3 subagent  
• [vocal-and-instrumental-stems] – The 2 MP3 URLs of the vocal and instrumental stems  
• [bpm-value] – The BPM (tempo) of the original audio track

E) SUBAGENT TASK SUMMARY  
1. Subagent receives the three inputs ([youtube-mp3-file], [vocal-and-instrumental-stems], [bpm-value]).  
2. It invokes Skill #223 (Powerful LLM Prompt-to-Text Response) and constructs a prompt that includes all three pieces of data.  
3. Skill #223 outputs a concise summary in text form, referencing the MP3 file URL, the two stem URLs, and the BPM value.  
4. This final text string is returned as the subagent’s final output, named [agent-summary].

F) SILOS  
• SINGLE SILO – Summarize Audio Data with an LLM  
  1) Input: [youtube-mp3-file], [vocal-and-instrumental-stems], [bpm-value]  
  2) Action Step: Skill #223 – Use a structured prompt to craft a text overview of the MP3 link, the two extracted stems, and the BPM.  
  3) Output: [agent-summary] (the final text summary).

4 Template & Links

Expand Flow

5 Template & Links

Expand Flow

A) SUBAGENT SUMMARY:  
This subagent’s purpose is to take the final data (original MP3 URL, separated stems, and BPM) and output a JSON-like text response that neatly packages all of those results together.

B) FINAL TASK OUTPUT:  
A single JSON-style text output containing the original MP3 URL, the instrumental stem URL, the vocal stem URL, and the BPM value. For example:  
{
  "originalMp3": "<URL>",
  "instrumentalMp3": "<URL>",
  "vocalMp3": "<URL>",
  "bpm": 120
}

C) SUBAGENT INPUT:  
• Original MP3 URL  
• Instrumental MP3 URL  
• Vocal MP3 URL  
• BPM value

E) SUBAGENT TASK SUMMARY:  
1) Receive the necessary data (original MP3, instrumental, vocal, BPM).  
2) Pass these inputs into Skill #223 ("Powerful LLM Prompt-to-Text Response") with instructions to format them as a JSON-like text object.  
3) Output the JSON-like string that includes the original MP3 URL, the instrumental stem URL, the vocal stem URL, and the BPM.

┌─────────────┐  
   │ Subagent #6 │  
   └─────────────┘  
         │  
         ▼  
   [Inputs: originalMp3URL, instrumentalMp3URL, vocalMp3URL, bpmValue]  
         │  
         ▼  
   Step 1: #223 - "Powerful LLM Prompt-to-Text Response"  
         (Prompt: “Output a valid JSON with keys: originalMp3, instrumentalMp3, vocalMp3, bpm…”)  
         └─> Output: JSON as text string  
         │  
         ▼  
   [Final Subagent Output: JSON-like text with the above fields]

F) SILOS:  
No complex silos are required. This subagent simply:  
• Gathers the final URLs and BPM value.  
• Uses one skill (#223) to generate an easy-to-return JSON text response.  
• Returns the JSON text as the final subagent output.