菜单
Use when transcribing audio to text via the local whisper.cpp server (large-v3, CUDA) — POST an audio file, get JSON back.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Anthropic API rate limit handling - retry logic, backoff, throttling for batch workloads against Claude models
Use when building an automated test → issue → fix loop with Claude Code and GitHub issues — overnight auto-fixing, regression loops, self-healing CI.
Use when creating, editing, publishing, or deleting posts on Cyril's Workshop blog or the steponnopets.net devblog.
Use when writing or contributing a boofuzz network-protocol fuzzer in this repo — layout, formatting rules, and reading results.
Use when a task needs real-time control of a connected browser via the Browser Bridge Broker — submit JS jobs over HTTP that browsers eval and return.
Use when training a character LoRA (Chroma/Flux or Pony/SDXL) on a RunPod GPU and wiring it into the ComfyUI + pony_web render stack.
| name | Whisper-Transcription |
| description | Use when transcribing audio to text via the local whisper.cpp server (large-v3, CUDA) — POST an audio file, get JSON back. |
Local speech-to-text via whisper.cpp with GPU acceleration on port 5555. Accepts WAV/MP3 and other common formats (16kHz mono WAV is optimal) as multipart POST.
curl -X POST http://localhost:5555/inference -F "file=@audio.wav"
Returns JSON {"text": ...}.
~/whisper.cpp/build/bin/whisper-server~/whisper.cpp/models/ggml-large-v3.binwhisper-server systemd service (systemctl status whisper-server, logs via journalctl -u whisper-server -f).Systemd unit at /etc/systemd/system/whisper-server.service:
[Unit]
Description=Whisper.cpp Transcription Server
After=network.target
[Service]
Type=simple
User=matt
WorkingDirectory=/home/matt/whisper.cpp/build
ExecStart=/home/matt/whisper.cpp/build/bin/whisper-server \
-m /home/matt/whisper.cpp/models/ggml-large-v3.bin \
-l en \
--port 5555 \
--host 0.0.0.0 \
--threads 4
Restart=always
RestartSec=10
[Install]
WantedBy=multi-user.target
~2–4x real-time; the first request after startup is slow (model load). large-v3 uses ≈ 6GB VRAM on the shared RTX 3090 — on OOM, wait for other GPU services to unload (see the Vram-GPU-OOM skill for the retry/signaling pattern).