trycua · JasonCaoGit · Sep 13, 2025 · Sep 13, 2025 · Sep 13, 2025 · Sep 13, 2025
diff --git a/.tool-versions b/.tool-versions
@@ -0,0 +1 @@
+python 3.11.7
diff --git a/bandlab_clean.mid b/bandlab_clean.mid
diff --git a/bandlab_cool_groove.mid b/bandlab_cool_groove.mid
diff --git a/bandlab_drums.mid b/bandlab_drums.mid
diff --git a/bandlab_groove.mid b/bandlab_groove.mid
diff --git a/clap.wav b/clap.wav
diff --git a/cua.py b/cua.py
@@ -0,0 +1,87 @@
+from computer import Computer
+import asyncio
+import logging
+from pathlib import Path
+from agent import ComputerAgent
+
+
+async def computer_use_agent(midi_file='./hum_basic_pitch.mid', instrument='piano'):
+    """
+    Main async function that handles the computer automation.
+    This function:
+    1. Reads the MIDI file created from your humming
+    2. Connects to a cloud computer environment  
+    3. Uploads the MIDI file to the remote computer
+    """
+    # Read the MIDI file that was generated from your humming
+    midi_file = open(midi_file, "rb")
+    content = midi_file.read()
+    midi_file.close()
+
+    print("📁 MIDI file loaded successfully")
+    print(f"📊 File size: {len(content)} bytes")
+
+    # Connect to the cloud computer environment
+    async with Computer(
+        os_type="linux",                    # Using Linux environment
+        provider_type="cloud",              # Cloud-based computer
+        name="l-linux-x318jdeu72",         # Specific instance name
+        api_key=""
+    ) as computer:
+        print("🖥️ Connected to cloud computer")
+
+        # Start the computer interface
+        await computer.run()
+        print("▶️ Computer interface started")
+
+        # Upload the MIDI file to the remote computer
+        await computer.interface.write_bytes("~/Downloads/midi-file-name.midi", content)
+        print("✅ MIDI file uploaded to ~/Downloads/midi-file-name.midi")
+        midi_name = "midi-file-name.midi"
+
+
+        agent = ComputerAgent(
+            model="anthropic/claude-opus-4-20250514",
+            tools=[computer],
+            max_trajectory_budget=5.0
+        )
+        tasks = [f"""
+You are inside BandLab Studio in Firefox on Linux.
+
+Goal: Import the MIDI file and play it with the chosen instrument.
+
+FILE TO IMPORT: "~/Downloads/midi-file-name.midi"
+FILE NAME ONLY: 
+INSTRUMENT: {instrument}
+
+Do the following step by step:
+
+1. Click the dashed box in the timeline that says “Drop a loop or an audio/MIDI/video file”.
+   - This should open a file upload dialog.
+
+2. In the file dialog:
+   - Click “Downloads” in the sidebar.
+   - Find and double-click “{midi_name}”.
+   - If it’s not visible, type “{midi_name}” into the filename field and press Enter.
+   - Wait until the MIDI region appears on the timeline.
+   - Take a screenshot.
+
+Rules:
+- Always interact inside BandLab, not the browser’s URL bar.
+- Use precise clicks; scroll if needed.
+"""]
+
+        for i, task in enumerate(tasks):
+            print(f"\nExecuting task {i}/{len(tasks)}: {task}")
+            async for result in agent.run(task):
+                print(result)
+            print(f"\n✅ Task {i+1}/{len(tasks)} completed: {task}")
+
+
+
+
+
+# This ensures the main function runs when the script is executed
+# It properly handles the async/await syntax
+if __name__ == "__main__":
+    asyncio.run(computer_use_agent())
diff --git a/drum_groove.mid b/drum_groove.mid
diff --git a/drum_groove_1757801992.mid b/drum_groove_1757801992.mid
diff --git a/drum_midi.py b/drum_midi.py
@@ -0,0 +1,65 @@
+import librosa
+import pretty_midi
+import sounddevice as sd
+import soundfile as sf
+
+def record_claps(filename="clap.wav", duration=8, samplerate=44100):
+    """Record clapping audio from mic."""
+    print(f"🎙️ Recording {duration} seconds... Clap steadily (like 1-2-3-4)!")
+    audio = sd.rec(int(duration * samplerate), samplerate=samplerate, channels=1, dtype="float32")
+    sd.wait()
+    sf.write(filename, audio, samplerate)
+    print(f"✅ Saved recording to {filename}")
+    return filename
+
+def clap_to_drum_groove(wav_file, out_midi="drum_groove.mid"):
+    """Use clapping as tempo guide, generate a clean 4/4 drum groove."""
+    # Load clap audio
+    y, sr = librosa.load(wav_file)
+
+    # Estimate tempo
+    tempo, _ = librosa.beat.beat_track(y=y, sr=sr)
+    if hasattr(tempo, "__len__"):  # sometimes returns array
+        tempo = float(tempo[0])
+    else:
+        tempo = float(tempo)
+
+    print(f"🎵 Detected tempo: {tempo:.1f} BPM")
+
+    # MIDI setup
+    pm = pretty_midi.PrettyMIDI(initial_tempo=tempo)
+    drums = pretty_midi.Instrument(program=0, is_drum=True)
+
+    beat_len = 60.0 / tempo   # seconds per beat
+    num_bars = 8              # how many bars to generate
+    total_beats = num_bars * 4
+
+    for i in range(total_beats):
+        t = i * beat_len
+        beat = i % 4
+
+        # Kick on 1 & 3
+        if beat == 0 or beat == 2:
+            drums.notes.append(pretty_midi.Note(110, 36, t, t+0.1))  # Kick drum
+
+        # Snare on 2 & 4
+        if beat == 1 or beat == 3:
+            drums.notes.append(pretty_midi.Note(110, 38, t, t+0.1))  # Snare drum
+
+        # Hi-hats every 8th note
+        for off in [0.0, 0.5]:
+            drums.notes.append(pretty_midi.Note(85, 42, t + off*beat_len, t + off*beat_len + 0.05))
+
+    pm.instruments.append(drums)
+    pm.write(out_midi)
+    print(f"🥁 Drum groove saved as: {out_midi}")
+    return out_midi
+
+if __name__ == "__main__":
+    # Step 1: Record your claps (used to detect tempo)
+    wav = record_claps("clap.wav", duration=8)
+
+    # Step 2: Generate structured groove MIDI
+    midi_file = clap_to_drum_groove(wav, "drum_groove.mid")
+
+    print("✅ Import 'drum_groove.mid' into BandLab, assign a drum kit, and enjoy a real groove!")
diff --git a/hum.py b/hum.py
@@ -0,0 +1,49 @@
+import sounddevice as sd
+import soundfile as sf
+import numpy as np
+import time
+def record_hum(output_file="hum.wav", instrument="piano", samplerate=44100, channels=1, silence_thresh=0.01, min_silence_len=2):
+    time.sleep(3)
+    print("🎤 Ready. Start humming...")
+
+    recording = []
+    silence_counter = 0
+    chunk_size = int(samplerate * 0.1)  # 100ms chunks
+
+    with sd.InputStream(samplerate=samplerate, channels=channels) as stream:
+        while True:
+            chunk, _ = stream.read(chunk_size)
+            chunk = chunk.copy()
+            recording.append(chunk)
+
+            volume = np.abs(chunk).mean()
+            if volume < silence_thresh:
+                silence_counter += 1
+            else:
+                silence_counter = 0
+
+            if silence_counter > min_silence_len * 10:  # e.g. 2s silence
+                break
+
+    recording = np.concatenate(recording, axis=0)
+    sf.write(output_file, recording, samplerate)
+    print(f"✅ Saved recording to {output_file}")
+
+    # Route to appropriate MIDI generator based on instrument
+    print(f"🎵 Processing for instrument: {instrument}")
+
+    if instrument.lower() in ["drums", "drum", "percussion"]:
+        # Use drum_midi.py for drum processing
+        print("🥁 Using drum MIDI generator...")
+        from drum_midi import clap_to_drum_groove
+        midi_file = clap_to_drum_groove(output_file, f"drum_groove_{int(time.time())}.mid")
+        return midi_file
+    else:
+        # Use midi.py for melodic instruments (piano, guitar, etc.)
+        print(f"🎹 Using melodic MIDI generator for {instrument}...")
+        from midi import convert_to_midi
+        midi_file = convert_to_midi(output_file, ".")
+        return midi_file
+
+if __name__ == "__main__":
+    record_hum()
diff --git a/hum.wav b/hum.wav
diff --git a/hum_1757800363_basic_pitch.csv b/hum_1757800363_basic_pitch.csv
@@ -0,0 +1,4 @@
+start_time_s,end_time_s,pitch_midi,velocity,pitch_bend
+8.840328344671201,9.014478004535148,60,70,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2
+3.9602861678004535,4.100889795918367,59,58,5,2,1,1,1,1,1,1,1,1,0,0,0
+1.1029478458049886,1.242267573696145,54,56,1,1,1,1,1,1,1,1,1,1,1,2
diff --git a/hum_1757800363_basic_pitch.mid b/hum_1757800363_basic_pitch.mid
diff --git a/hum_1757801953.wav b/hum_1757801953.wav
diff --git a/hum_1757801953_1757801960_basic_pitch.csv b/hum_1757801953_1757801960_basic_pitch.csv
@@ -0,0 +1 @@
+start_time_s,end_time_s,pitch_midi,velocity,pitch_bend
diff --git a/hum_1757801953_1757801960_basic_pitch.mid b/hum_1757801953_1757801960_basic_pitch.mid
diff --git a/hum_1757801981.wav b/hum_1757801981.wav
diff --git a/hum_basic_pitch.csv b/hum_basic_pitch.csv
@@ -0,0 +1,21 @@
+start_time_s,end_time_s,pitch_midi,velocity,pitch_bend
+6.296459410430839,6.482219047619048,47,71,2,2,2,2,1,0,0,0,0,1,1,1,1,2,2,2
+5.923656235827664,6.064259863945578,47,62,0,1,0,0,0,0,0,0,0,0,0,0,0
+5.71467664399093,5.923656235827664,47,47,2,2,0,0,0,2,2,-1,-1,1,1,1,1,1,1,1,1,0
+5.308327437641723,5.645016780045351,50,83,0,0,1,1,1,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
+5.1690077097505664,5.308327437641723,50,55,0,0,0,0,0,-1,-1,-1,-1,-1,-1,0
+4.960028117913832,5.110957823129252,52,57,0,0,1,1,1,2,3,3,3,3,3,2,2
+4.623338775510204,4.785878458049886,53,86,2,2,1,1,1,1,1,1,1,1,1,1,1,1
+4.228599546485261,4.623338775510204,53,67,1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,2,2
+3.9138462585034013,4.042839909297052,52,66,2,2,2,2,2,2,2,2,2,2,2,3
+3.3449573696145123,3.820966439909297,51,67,3,2,2,0,0,0,0,0,0,0,0,0,0,0,-1,-1,-1,-1,-1,-1,-1,-1,-1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
+2.8689482993197277,3.252077551020408,49,90,0,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,1,1,1,1,2,2,2,2,2,2,2,2,2,2,2,3,3
+2.2768394557823126,2.590308843537415,54,56,-1,-1,0,0,0,0,0,0,0,0,0,-1,-1,-1,-1,-1,-1,-1,-1,0,0,0,0,0,0,0,-1
+2.091079818594104,2.2768394557823126,53,98,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1
+1.8808163265306121,2.021419954648526,56,78,0,0,0,0,0,0,0,0,0,0,0,-1,-1
+1.6718367346938776,1.8808163265306121,54,85,2,2,1,1,1,1,1,1,1,1,1,1,1,1,2,2,1,1
+1.509297052154195,1.6950566893424037,56,90,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,-4
+1.1145578231292517,1.509297052154195,56,95,0,1,1,1,1,1,1,1,0,0,0,0,0,0,0,0,0,0,0,0,0,1,1,1,1,1,1,1,1,1,1,1,1,1
+0.9520181405895691,1.1145578231292517,56,103,-1,0,0,0,1,1,1,1,1,1,1,1,1,1
+0.8010884353741496,0.9520181405895691,55,99,1,1,1,1,1,1,1,1,1,1,1,1,2
+4.042839909297052,4.216989569160997,53,81,0,0,0,0,0,1,1,1,1,1,1,1,0,0,0
diff --git a/hum_basic_pitch.mid b/hum_basic_pitch.mid
diff --git a/instrument.txt b/instrument.txt
@@ -0,0 +1 @@
+Drums
diff --git a/instrument.wav b/instrument.wav
diff --git a/listen.js b/listen.js
@@ -0,0 +1,35 @@
+// Import the server SDK which works with Node.js (not browser-only like @vapi-ai/web)
+import { VapiClient } from '@vapi-ai/server-sdk';
+
+// Initialize Vapi client with your API key
+const vapi = new VapiClient({
+  apiKey: '1910ede1-e8f5-4c5e-aba2-deb03c53c9db'
+});
+
+console.log('🤖 Vapi server SDK initialized successfully!');
+
+// Async function to handle API calls
+async function testVapi() {
+  // Example: Create a call using the server SDK
+  // This is different from the web SDK - server SDK is for managing calls, not direct voice interaction
+  try {
+    const call = await vapi.calls.create({
+      assistant: {
+        id: 'a6210139-4abf-4ed8-b1a5-0996953045b3'
+      },
+      // For server SDK, you typically need to specify how to connect (phone, etc.)
+      customer: {
+        number: '+1234567890' // This would be the customer's phone number
+      }
+    });
+
+    console.log('📞 Call created:', call.id);
+  } catch (error) {
+    console.log('ℹ️ Note: Server SDK is for managing calls, not direct voice interaction');
+    console.log('For direct voice interaction, you need a web browser environment');
+    console.log('Error details:', error.message);
+  }
+}
+
+// Run the test
+testVapi();
Original file line number	Diff line number	Diff line change
		@@ -0,0 +1 @@
		start_time_s,end_time_s,pitch_midi,velocity,pitch_bend