fun-asr-nanoFun-ASR-Nano 基于 sherpa-onnx 引擎的本地语音识别,完全离线运行。当用户需要转录音频文件时触发。
Install via ClawdBot CLI:
clawdbot install pengzhendong/fun-asr-nanoGrade Fair — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Generated Oct 6, 2026
Reporters can transcribe interviews, press conferences, and field recordings locally without uploading sensitive audio to cloud services. The offline nature ensures source protection and confidentiality, while support for multiple Chinese dialects helps cover regional stories accurately.
Teams can automatically generate text records of meetings, including those with participants speaking different dialects or accents. The local processing eliminates privacy concerns and works even in low-connectivity environments, making it ideal for hybrid work setups.
Law firms and courts can transcribe witness testimonies, hearings, and client interviews on-premises, maintaining attorney-client privilege and data sovereignty. Dialect support ensures accurate capture of regional speech patterns common in legal proceedings.
Healthcare providers can convert patient consultations into searchable text without risking patient data exposure to third-party servers. The offline operation complies with strict privacy regulations, and dialect recognition aids in understanding elderly or regional patients.
Music producers and artists can transcribe lyrics and rap verses accurately, leveraging the model's specialized support for song lyrics and rap speech. This speeds up publishing metadata, subtitling, and copyright documentation while preserving creative privacy.
The core Fun-ASR-Nano skill is freely available under an open-source license, with paid enterprise support plans offering priority bug fixes, custom dialect model fine-tuning, and deployment assistance. This fosters community adoption while monetizing large organizations that require SLA-backed reliability.
A hosted version of the offline engine is offered as a subscription-based API, allowing users to upload audio and receive transcriptions via a web interface or REST API. While the core engine runs locally, the SaaS platform adds value through batch processing, collaboration tools, and storage.
The skill is packaged as a lightweight SDK that hardware manufacturers (e.g., voice recorders, smart speakers) and software vendors (e.g., video editors, note-taking apps) can embed into their products. Licensing fees are charged per device or per application sold.
💬 Integration Tip
Install dependencies via pip and run the provided CLI script; for custom applications, import the Python module directly and handle audio formats supported by soundfile. Ensure models are downloaded locally before offline use.
Scored Oct 6, 2026
Check Antigravity account quotas for Claude and Gemini models. Shows remaining quota and reset times with ban detection.
使用豆包(火山引擎)语音合成大模型 API 将文本转换为语音音频文件。支持声音复刻音色(S_ 开头的音色ID)和官方预置音色。当用户要求"语音合成"、"文字转语音"、"TTS"、"朗读文本"、"生成语音"、"用我的声音读"、"豆包语音"、"声音复刻合成"等相关请求时,务必使用此 skill。即使用户只是说"帮我把...
Intelligent model routing for sub-agent task delegation. Choose the optimal model based on task complexity, cost, and capability requirements. Reduces costs...
自动生成科技新闻摘要。从多个来源(RSS、Twitter、GitHub、Web Search)抓取科技新闻,整合后生成摘要。
让 AI 代理根据对话内容自动选择最合适的模型。四层识别(系统过滤→关键词→指示词→语义相似度),四池架构(高速/智能/人文/代理),五分支路由,全自动 Fallback 回路。支持 trigger_groups_all 非连续词组命中。
OpenAI API integration — chat completions, embeddings, image generation, audio transcription, file management, fine-tuning, and assistants via the OpenAI RES...