song-recognition

What This Skill Does

The song-recognition skill provides a powerful interface to iFlytek's Query By ACRCloud engine, enabling the OpenClaw agent to identify musical tracks through audio input. Whether a user provides a recorded audio file or captures a singing melody, the skill processes the acoustic data against a vast database to deliver precise track information. By leveraging advanced audio fingerprinting technology, it matches humming, singing, or instrumental snippets to their original studio recordings, providing essential metadata such as the song title, performing artist, album details, and a confidence score for each match.

Installation

To install this skill, use the ClawKit CLI within your terminal: clawhub install openclaw/skills/skills/dzy-1026/xfyun-song-rec

Ensure your development environment has Python installed and correctly configured in your system PATH. Once installed, you must provide your iFlytek credentials. This can be done by setting environment variables in your shell (e.g., XF_SONG_APP_ID, XF_SONG_API_KEY, and XF_SONG_API_SECRET) or by adding these keys to your ~/.openclaw/openclaw.json configuration file. A valid network connection is required to communicate with the iFlytek recognition servers.

Use Cases

• Music Discovery: Instantly identify a catchy tune you heard on the radio or in a public space by recording a short clip.
• Karaoke Enhancement: Use the skill to verify song titles for library management or to build automated karaoke playlists.
• Content Moderation: Identify background music in user-uploaded audio to assist in copyright detection workflows.
• Education: Assist music students in identifying pieces by melody, even if they only have a rough vocal recording.

Example Prompts

1. "OpenClaw, identify the song in this audio clip: ./recordings/mystery_hum.wav"
2. "What song is being played in this file? Analyze hummingbird.wav and return the artist and album info."
3. "Please run the song recognition skill on my latest recording and save the artist metadata to a text file."

Tips & Limitations

• Audio Quality: Clarity is crucial. Ensure the audio contains a distinct, recognizable melody. Background noise can significantly degrade the match accuracy.
• Format: The skill currently supports MP3 files. Aim for a sample rate of 16000Hz using lame encoding for optimal results.
• Duration: For best performance, provide audio clips between 5 and 30 seconds. Extremely short or very long clips may cause processing errors.
• Synchronous Nature: The skill operates synchronously, meaning the agent will pause to process the identification; ensure your automation flow accounts for this response time.