Kokoro-FastAPI is a Dockerized FastAPI service that turns text into speech using the Kokoro-82M model through an OpenAI-compatible API. It is used by developers and self-hosters who need multilingual, multi-speaker speech generation, captions, phonemes, or an optional web interface.
Borrowing it
Nothing to install: this file belongs to remsky/Kokoro-FastAPI. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/remsky/Kokoro-FastAPI/master/.claude/skills/readme-benchmarks/SKILL.mdgit clone --depth 1 https://github.com/remsky/Kokoro-FastAPIWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/remsky/kokoro-fastapi/readme-benchmarks)<a href="https://agentmods.dev/skills/remsky/kokoro-fastapi/readme-benchmarks"><img src="https://agentmods.dev/badge/skills/remsky/kokoro-fastapi/readme-benchmarks.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00052 | $0.01753 |
| Opus 5 | $0.00026 | $0.00877 |
| Sonnet 5 | $0.00010 | $0.00351 |
| Haiku 4.5 | $0.00005 | $0.00175 |
Grade A, and why
readme-benchmarks scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 8d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Benchmarks + README plots
Three suites under examples/assorted_checks/ feed the README:
test_transcription/- synth with a running server, transcribe with faster-whisper, report WER/CER. Short, multilingual, and long-form.benchmarks/- RTF (processing time vs tokens) and first-token latency/timeline plots. These feed the README performance images.test_dialogue/- multi-speaker functional checks plus the turn-length / text-length throughput sweeps. Run commands and flags live intest_dialogue/README.md; bench then plot withplot_dialogue_bench.py.
Prereqs
- A Kokoro server on
:8880. GPU and CPU docker images both bind 8880, so swap images, never run both. SetKOKORO_DEVICE/BENCH_PREFIXto match whichever is up. - Warm the server first. cuDNN autotune cold-start inflates the first GPU run (heavy voices ~1.2s vs ~0.3s warm). Hit a couple voices (e.g.
af_heart,zf_xiaobei) before capturing, or discard the first pass. examples/has its own uv venv. Run everything from insideexamples/(cd examples), not the root.venv. The first-token script also needsexamples/as cwd for its audio path.
cd examples
uv sync --extra transcription --extra transcription-gpu --extra benchmarks
Drop --extra transcription-gpu to skip the ~1.2 GB cuDNN/cuBLAS download and transcribe on CPU.
Run (from examples/)
# short English per-voice sanity (base.en, WER)
KOKORO_DEVICE=gpu uv run python assorted_checks/test_transcription/test_transcription.py
# multilingual (small model, CER for ja/zh)
KOKORO_DEVICE=gpu uv run python assorted_checks/test_transcription/test_transcription_multilingual.py
# long-form book roundtrip (the baseline). see BASELINE.md
LONGFORM_CHARS=65000 WHISPER_DEVICE=cuda KOKORO_DEVICE=gpu \
uv run python assorted_checks/test_transcription/test_long_form.py
# RTF plots
BENCH_PREFIX=gpu uv run python assorted_checks/benchmarks/benchmark_tts_rtf.py
# first-token latency/timeline plots
BENCH_PREFIX=gpu uv run python assorted_checks/benchmarks/benchmark_first_token_stream_unified.py
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 8d ago First seen · 92 lines · 52 tokens per session scan A 20440e309fbb
readme-benchmarks is a skill published in the GitHub repository remsky/Kokoro-FastAPI (5,414 stars, last pushed yesterday), licensed Apache-2.0. It adds 52 tokens to every session and 1,753 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
voice-tts
Text-to-speech output for voice responses. Uses pyttsx3 for local TTS on supported systems.
torchforge-rl-training
Provides guidance for PyTorch-native agentic RL using torchforge, Meta's library separating infra from algorithms. Use when you want clean RL abstractions, easy algorithm experimentation, or scalable training with Monarch and TorchTitan.
datamodel-code-generator
Use this skill when the user wants Python data models, Pydantic models, dataclasses, TypedDicts, msgspec structs, or type-safe Python classes generated from OpenAPI, AsyncAPI, JSON Schema, GraphQL, JSON/YAML/CSV sample data, MCP tool schemas, Protocol Buffers, XML Schema, Apache Avro, or existing Python model objects.…
experiment-tracking-swanlab
Provides guidance for experiment tracking with SwanLab. Use when you need open-source run tracking, local or self-hosted dashboards, and lightweight media logging for ML workflows.
video-script
A Chinese-language video editing and narration workflow that plans the story, selects clips, assigns visuals and sound, writes timestamped narration, and validates the result.
oss-maintainer
Run an open-source project's issue/PR/release loop like a careful human maintainer — triage to root cause, absorb community PRs before duplicating them, gate every merge, ship honest releases, and thank the people doing your QA for free.