elevenlabs_skill

Overview

This skill provides ready-to-use curl examples for ElevenLabs AI text-to-speech. It helps you list voices, inspect models, generate high-quality audio files, and stream real-time audio using the ElevenLabs API. Examples assume the ELEVENLABS_API_KEY environment variable is set.

How this skill works

The skill uses direct HTTP calls to the ElevenLabs v1 API via curl, passing the API key in the xi-api-key header. It shows endpoints for listing voices and models, fetching voice details, submitting text-to-speech requests (file or streamed), and checking subscription usage. Request payloads and common response handling with jq are demonstrated.

When to use it

• Generate realistic voiceovers for videos, podcasts, or accessibility features
• Stream audio for live playback or low-latency applications
• Explore and select voices or models available to your account
• Save synthesized speech to files in MP3 or PCM formats
• Monitor subscription usage and character limits

Best practices

• Store your API key in ELEVENLABS_API_KEY and use bash -c when piping to avoid environment variable clearing
• Pick the appropriate model: eleven_multilingual_v2 for quality, eleven_flash_v2_5 for low latency
• Cache frequently used voice IDs to reduce repeated list calls
• Tune voice_settings (stability and similarity_boost) to balance consistency and expressiveness
• Use the streaming endpoint for long text or real-time UX to reduce latency

Example use cases

• Batch-generate narration MP3s for an e-learning course using a chosen voice ID
• Stream short prompts in a chatbot for instant audible replies with eleven_flash_v2_5
• Clone and test a custom voice, then adjust similarity_boost for final output
• Export PCM audio for further processing in audio pipelines or ASR systems
• Automate usage checks by calling the subscription endpoint to avoid overruns