Update to VoiceLab

{
  "description": "Get started with VoiceLab speech AI tools for TTS, STT, Meeting AI, file upload, voice isolation, and LLM.",
  "included_files": [],
  "name": "get-started",
  "skill_md_contents": "---\nname: get-started\ndescription: Get started with VoiceLab speech AI tools for TTS, STT, Meeting AI, file upload, voice isolation, and LLM.\n---\n\n# Getting Started with VoiceLab\n\nWelcome to VoiceLab! This guide will help you start using VoiceLab's speech AI capabilities.\n\n## What VoiceLab Does\n\nVoiceLab provides production-grade speech AI tools for agents:\n\n- **Text-to-Speech (TTS)**: Generate natural speech in Uzbek, Russian, and English\n- **Speech-to-Text (STT)**: Transcribe audio with word-level timing and speaker labels\n- **Meeting AI**: Upload a meeting recording and get a transcript plus AI notes\n- **File upload**: Attach an audio file in ChatGPT, or pass a public HTTPS link (`audio_url`)\n- **Voice Isolation**: Remove background noise with optional speech restoration\n- **LLM Completions**: Chat with Aisha models for Uzbek language understanding\n\n## Authentication\n\nVoiceLab uses Bearer token authentication. You'll need a VoiceLab MCP gateway Bearer token when connecting.\n\nGet your API key at [voicelab.uz](https://voicelab.uz).\n\n## Quick Start Examples\n\n### Generate Speech\n\n```\nGenerate Uzbek speech for: \"Salom, dunyo!\"\n```\n\nThe agent will:\n1. Call `list_voices` to find available Uzbek voices\n2. Call `text_to_speech` with your text\n3. Return base64-encoded WAV audio\n\n### Transcribe Audio\n\n```\nTranscribe this meeting recording with speaker labels.\n[Attach audio file]\n```\n\nThe agent will:\n1. Call `speech_to_text` with your audio (base64-encoded)\n2. Poll `get_transcription` until processing completes\n3. Return transcription with word timing and speaker labels\n\n### Meeting Notes\n\n```\nSummarize this meeting recording.\n[Attach audio file]\n```\n\nThe agent will:\n1. Call `create_meeting_from_file` (attached file) or `create_meeting` with `audio_url`\n2. Poll `get_meeting` until status is `ready`, `transcribed`, or `failed`\n3. Return transcript and notes (use `regenerate_meeting_notes` if notes failed)\n\nNote: `get_meeting_translations` is not available with API keys.\n\n### Remove Noise\n\n```\nClean up this phone call recording by removing background noise.\n[Attach audio file]\n```\n\nThe agent will:\n1. Call `isolate_voice` with your audio\n2. Poll `get_isolation` until processing completes\n3. Return isolated audio URL\n\n## Key Concepts\n\n### Audio Format\n- **Input**: Attach a file in ChatGPT, give a public HTTPS `audio_url`, or send small files as base64 (under ~7 MB)\n- **Output**: Generated audio is returned as base64-encoded WAV (24kHz)\n- **Supported formats**: MP3, WAV, M4A, OGG, WebM, FLAC\n\n### Async Operations\nSTT and voice isolation are asynchronous:\n1. Submit job (returns job ID immediately)\n2. Poll status endpoint until `status: \"completed\"`\n3. Access results from completed response\n\n### Languages\n- **Uzbek (uz)**: Full TTS and STT support\n- **Russian (ru)**: Full TTS and STT support  \n- **English (en)**: Full TTS and STT support\n\n### Idempotency\nAll operations support idempotency keys. The agent auto-generates UUIDs if you don't provide them. Retrying with the same key returns the cached result without reprocessing.\n\n## Common Workflows\n\n### TTS Workflow\n```\nUser → list_voices (optional) → text_to_speech → audio_base64\n```\n\n### STT Workflow\n```\nUser → speech_to_text → get_transcription (poll) → transcription result\n```\n\n### Voice Isolation Workflow\n```\nUser → isolate_voice → get_isolation (poll) → audio URLs\n```\n\n## Tips\n\n1. **Always list voices first** when generating speech in a new language\n2. **Check supported languages** with `list_tts_languages` before TTS\n3. **Enable speaker labels** for multi-speaker transcriptions with `include_speakers: true`\n4. **Use speech restoration** for low-quality audio with `speech_restoration: true`\n5. **Export transcriptions** as SRT or VTT for video subtitles\n\n## Support\n\n- Documentation: [docs.voicelab.uz](https://docs.voicelab.uz)\n- Website: [voicelab.uz](https://voicelab.uz)\n- Contact: elzodxon@gmail.com\n\n## What's Next\n\nTry these commands:\n- \"Show me all available Uzbek voices\"\n- \"Generate Russian speech for: Привет, мир!\"\n- \"Transcribe this audio with word timing\"\n- \"Remove background noise and restore speech quality\"\n"
}