LocalAI/backend/python/kokoro
Ettore Di Giacinto 2de30440fe
fix(l4t-12): use pip to install python deps (#7967)
* fix: install only torch/torchvision from jetson index

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* fix: use pip for l4t-12

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* Revert "fix: install only torch/torchvision from jetson index"

This reverts commit 2d2b020078

* chatterbox needs wheel

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

---------

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
2026-01-11 00:21:32 +01:00
..
backend.py feat: return complete audio for kokoro (#6842) 2025-10-28 08:49:18 +01:00
install.sh fix(l4t-12): use pip to install python deps (#7967) 2026-01-11 00:21:32 +01:00
Makefile feat(mlx): add mlx backend (#6049) 2025-08-22 08:42:29 +02:00
README.md feat(kokoro): complete kokoro integration (#5978) 2025-08-06 15:23:29 +02:00
requirements-cpu.txt feat(kokoro): complete kokoro integration (#5978) 2025-08-06 15:23:29 +02:00
requirements-cublas12.txt feat(kokoro): complete kokoro integration (#5978) 2025-08-06 15:23:29 +02:00
requirements-cublas13.txt feat: add cuda13 images (#7404) 2025-12-02 14:24:35 +01:00
requirements-hipblas.txt chore: Update to Ubuntu24.04 (cont #7423) (#7769) 2026-01-06 15:26:42 +01:00
requirements-intel.txt feat(kokoro): complete kokoro integration (#5978) 2025-08-06 15:23:29 +02:00
requirements-l4t12.txt feat: add cuda13 images (#7404) 2025-12-02 14:24:35 +01:00
requirements.txt feat(kokoro): complete kokoro integration (#5978) 2025-08-06 15:23:29 +02:00
run.sh feat: Add backend gallery (#5607) 2025-06-15 14:56:52 +02:00
test.py feat(kokoro): complete kokoro integration (#5978) 2025-08-06 15:23:29 +02:00
test.sh feat: Add backend gallery (#5607) 2025-06-15 14:56:52 +02:00

Kokoro TTS Backend for LocalAI

This is a gRPC server backend for LocalAI that uses the Kokoro TTS pipeline.

Creating a separate environment for kokoro project

make kokoro

Testing the gRPC server

make test

Features

  • Lightweight TTS model with 82 million parameters
  • Apache-licensed weights
  • Fast and cost-efficient
  • Multi-language support
  • Multiple voice options