🔒 Verified source: github.com/hexgrad⏳ Pending test
📄🧭🔗📚🧭🗺️🔗
🧩 Teach this AI new skills
Unlock 8 new powers for your AI agent
Kokoro has no skill store we can verify. 7 popular cross-AI skills for agents like Claude Code and Cursor, mapped to what your AI agent can do after you install them · tap a power
Pre-check: open Task Manager → Performance to see RAM and GPU. Then run this in Command Prompt to see GPU name, driver version and VRAM. Also check free disk space.
nvidia-smi
Do you see your GPU name and "Driver Version"?
Install espeak-ng: the Windows .msi from GitHub releases (espeak-ng/espeak-ng).
Create speak.py with this code (for Hindi use lang_code='h' and voice hf_alpha).
from kokoro import KPipeline
import soundfile as sf
pipe = KPipeline(lang_code='a') # a = American English · h = Hindi
for i, (gs, ps, audio) in enumerate(pipe("Hello, this is Kokoro speaking.", voice='af_heart')):
sf.write(f'out_{i}.wav', audio, 24000)
Run — the first run downloads the model (~330 MB).
python speak.py
Does out_0.wav play?
Pre-check: Task Manager → Performance → GPU shows "AMD Radeon …" and dedicated GPU memory (VRAM). Update the driver with AMD Software: Adrenalin (amd.com).
Can you see the GPU name and VRAM?
Install espeak-ng: the Windows .msi from GitHub releases (espeak-ng/espeak-ng).
Create speak.py with this code (for Hindi use lang_code='h' and voice hf_alpha).
from kokoro import KPipeline
import soundfile as sf
pipe = KPipeline(lang_code='a') # a = American English · h = Hindi
for i, (gs, ps, audio) in enumerate(pipe("Hello, this is Kokoro speaking.", voice='af_heart')):
sf.write(f'out_{i}.wav', audio, 24000)
Run — the first run downloads the model (~330 MB).
python speak.py
Does out_0.wav play?
Pre-check: Apple menu → About This Mac. Note the Chip (Apple M-series or Intel), Memory (RAM) and macOS version.
Create speak.py with this code (for Hindi use lang_code='h' and voice hf_alpha).
from kokoro import KPipeline
import soundfile as sf
pipe = KPipeline(lang_code='a') # a = American English · h = Hindi
for i, (gs, ps, audio) in enumerate(pipe("Hello, this is Kokoro speaking.", voice='af_heart')):
sf.write(f'out_{i}.wav', audio, 24000)
Run — the first run downloads the model (~330 MB).
Create speak.py with this code (for Hindi use lang_code='h' and voice hf_alpha).
from kokoro import KPipeline
import soundfile as sf
pipe = KPipeline(lang_code='a') # a = American English · h = Hindi
for i, (gs, ps, audio) in enumerate(pipe("Hello, this is Kokoro speaking.", voice='af_heart')):
sf.write(f'out_{i}.wav', audio, 24000)
Run — the first run downloads the model (~330 MB).
python speak.py
Does out_0.wav play?
Pre-check: see your RAM (Task Manager / About This Mac / free -h). Small tools run without a GPU — just slower.
Know how much RAM you have?
Kokoro runs fine on CPU — follow your OS tab.
Sizes are approximate and can change on the official page.
Errors & fixes
pip: "Could not find a version that satisfies kokoro"
Your Python is 3.13 or 3.9 — kokoro needs 3.10–3.12.
espeak / phonemizer error
Install espeak-ng and open a new terminal.
"python is not recognized" / "command not found: python"
Python is missing or not on PATH. Install from python.org and tick "Add python.exe to PATH" on the first installer screen, then open a new terminal. On Mac/Linux use python3.
Download stuck / very slow / "connection reset"
Network or DNS issue. Restart the router, try a mobile hotspot, or set DNS to 1.1.1.1 / 8.8.8.8. Re-running the same command usually resumes the download.
Popular AI skills that work with tools like Kokoro 7 skills
Kokoro has no skill store we can verify. These cross-AI skills (MCP servers and Agent Skills) work in most AI agents — each card shows one example install; see all AIs for yours.
MCPPlugins
🛡 Stay safe
Read first
Official first — from the AI maker or the service’s own GitHub org / domain.
Read SKILL.md and any scripts before installing. Skills can run code with your permissions.
Never paste base64 or “curl … | sh” download-and-run lines from a skill.
Stars, license and last update come from GitHub (17 Nov 2026). Commands are copied from official docs and are not tested on a real machine yet. Access tags are our summary — always check the source. Skills for every AI →