Lightweight open-weights text-to-speech model for fast, natural voice generation.
Verified free allowance
Running Kokoro locally or via open-weight implementations has no provider rate limits or quotas; resource capacity depends solely on host hardware.
Kokoro is an efficient 82-million parameter open-weights text-to-speech model created by hexgrad. Licensed under Apache 2.0, it runs with high speed on consumer CPUs and GPUs, supports multiple languages and voices, and can be run locally or integrated into self-hosted APIs without request quotas.
An exceptional open-weights choice for developers and creators who want natural text-to-speech without subscription fees, usage limits, or proprietary licensing restrictions.
Paid plans: The model weights and code are free under Apache 2.0; optional commercial hosted API providers charge standard inference rates.
These links support the important free-offer claims on this page.
Kokoro is an 82M parameter TTS model released under the Apache 2.0 license with open weights and commercial rights.
Kokoro provides open-weights inference code for local and API-driven speech synthesis without provider-imposed rate limits.
