Creative Claw Cartesia Sonic

Read shared execution guidance once per task before using tools. It covers model discovery, optional cost checks, imports, and recovery.

Use speech/cartesia-sonic for fast, natural speech from a public stock voice or a saved consented Character voice. The Creative Claw model ID stays stable while the underlying Cartesia Sonic version can be upgraded.

Choose deliberately

Cartesia is a strong first choice for low latency, direct primary emotion control, and simple speed or volume adjustments. ElevenLabs v2 is a strong first choice for steady professional narration. ElevenLabs v3 is a strong first choice for acted delivery with inline audio tags.

Honor an explicit model choice. Never switch providers silently. When the first result misses the intended voice or performance, offer a controlled comparison using the same text, language, pacing, and output format.

Choose a voice

Call get_model_params({ model: "speech/cartesia-sonic" }) before generation. It returns the current curated Featured catalog with exact IDs, language, locale, gender, description, use cases, and the full Cartesia Voice Library URL.

Use exactly one selector: