Products

FunASR

AI-native

让语音,成为可用的信息。

funasr.com Source★ 2,754Video & Audio / Transcription & Dictation

AI Benchmark
Desktop
Mobile
Monthly visits
34.6K
Down20.2%MoMOct 2026
Public revenue
Undisclosed
No public revenue evidence yet
Entry price
Free
No monthly subscription pricing
Launched
6mo ago
2026-05-24 · Est.

Traffic

Last 12 months
Monthly visitsData through Oct 2026
Events3 marketing moves

Marketing moves

Launches, search & content, referrals, promotions
  • 【工具自荐】FunASR —— 比 Whisper 快 170 倍的开源语音识别,新增 Agent 友好 CLI (1 comments)Launch

    【工具自荐】FunASR —— 比 Whisper 快 170 倍的开源语音识别,新增 Agent 友好 CLI — ## 简介 [FunASR](https://github.com/modelscope/FunASR)(17,700+ stars)是阿里达摩院开源的工业级语音识别工具包,支持 50+ 种语言、说话人分离、情感识别。 **最近新增了 Agent 友好的 CLI**,一行命令即可转写音频: ```bash pip install funasr # 转写音频 funasr audio.wav # JSON 输出(方便 AI Agent 调用) funasr audio.wav --output-format json # 生成 SRT 字幕 funasr audio.wav -f srt --output-dir ./subs #…

    Jun 10 · 5mo agoSource

  • 【开源推荐】SenseVoice — 一个模型同时做语音识别+情感检测+声音事件分类 (1 comments)Launch

    【开源推荐】SenseVoice — 一个模型同时做语音识别+情感检测+声音事件分类 — ## 项目介绍 [SenseVoice](https://github.com/FunAudioLLM/SenseVoice)(8.2K Stars)是阿里通义实验室的多任务语音理解模型。 **一个 234MB 模型同时完成:** - 语音识别(ASR) - 语种检测(中/英/日/韩/粤) - 情感识别(开心/悲伤/愤怒/中性) - 声音事件检测(笑声/掌声/音乐/哭泣) ## 速度 - GPU: **170 倍实时**(10 秒音频只需 70ms) - CPU: **17 倍实时**(无需 GPU 也能用) ## 使用 ```python pip install funasr from funasr import AutoMod…

    May 24 · 6mo agoSource

  • 【开源推荐】FunASR 语音输入法 — 按快捷键说话,自动转文字粘贴到光标位置Launch

    【开源推荐】FunASR 语音输入法 — 按快捷键说话,自动转文字粘贴到光标位置 — ## 项目介绍 [FunASR](https://github.com/modelscope/FunASR)(16K+ Stars)是阿里通义实验室开源的工业级语音识别工具包。最近推出了桌面语音输入法工具: **按快捷键 → 说话 → 自动识别 → 粘贴到光标位置** 支持中英日韩粤 5 种语言,GPU 170 倍实时 / CPU 17 倍实时,完全本地运行。 ## 使用方式 ```bash pip install funasr sounddevice numpy pyperclip openai pynput # 启动语音服务 funasr-server --device cuda # 或 cpu # 启动语音输入法 pytho…

    May 24 · 6mo agoSource

Third-party reviews & case dossiers

Acquisition breakdowns, engineer reviews & open-source metrics
Ruanyf Weekly
2026-05-24Source
Ruanyf Weekly
2026-05-24Source
Ruanyf Weekly
2026-06-10Source

Open Source & Agent Ecosystem

GitHub · MCP / Skill
Last commit 2026-07-24
GitHub Stars
★ 2,754
Forks
252
License
See repository
Primary language
Go

Traffic sources

Oct 2026

Traffic breakdown

Channels

  • Organic Search49%
  • Direct31%
  • Referrals16%
  • Other4%

Countries

  • China73%
  • United States8%
  • Taiwan5%
  • Brazil5%
  • Vietnam4%

Search

Domain authority —Domain registered 2023-08-28

KeywordMonthly volume
funasr2,390
cpu 可以运行的 asr模型200
opensource repos that can make srt on video140
asr模型80
fun-asr40

Traffic data: Vaneform

Pricing

Pricing page not identified yet, or no public pricing listed.

Pages & changes

Tracked since 2026-08-01

Key pages

Weekly snapshots
funasr.com/ Snapshot 2026-11-04

Official blog & changelog

Recent articles & product updates
  • 2026-11-14BlogRead post

    长录音为什么会漏掉结尾?

    如果团队拿到了长录音的文本,却找不到最后一句,先别急着调大批次。沿着原文件、语音分段和返回结果定位,才能知道哪一步需要处理;有文本并不代表全程完整。 设想一段访谈:原文件里能听到最后一句,文本里却没有。这个现象还不能直接归因于模型。先确认读取包含尾部、分段保留了那句声音,再看客户端有没有保存完整结果。 下面只用一个明确的 SenseVoice 离线分段配置说明排查路径,不是所有漏尾问题的修复。静音、切分、推理和客户端异常需要分别核验。 调参之前,先查三个位置 1. 输入: 核

  • 2026-11-14BlogRead post

    把语音转写接进自己的 API

    如果应用只需要上传一段录音、取回文字,先把这一个来回走通。团队可以先在本机核对请求与返回值,再决定怎样承接真实流量,而不是一开始就复制整套云服务。 这里用一份获授权的 `audio.wav` :客户端把文件发给本机 SenseVoice 服务,再读取 JSON 的 `text` 字段。示例展示调用路径,不编造转写结果,也不比较速度。 本文使用 FunASR 1.4.15 包内的 `funasr-server` 。它只提供部分 OpenAI 转写接口,不是所有云功能和返回格式的

  • 2026-11-14BlogRead post

    接入 Transformers,先选对权重

    如果你已经在用 Transformers,接入语音识别时先确认权重属于这条接口。同一个模型名字,不代表工具包、服务端和原生 Python 能互换文件。 例如,文件下载完了,应用却提示“不认识模型类型”。先检查安装包有没有实现、权重是不是对应格式,不要急着把问题归为文件损坏。本篇选择官方 `-hf` 权重。 这条路径生成转写文本,不是现成服务。加载成功不等于拿到了字级时间戳、说话人分离或实时流式能力。 想直接试一段录音?查看 Transformers 5.17.0 快速开始 ,

  • 2026-11-14BlogRead post

    会议转写,怎样才算做好了? | FunASR

    团队准备交付会议纪要或字幕时,需要确认名字、决策和原声音能对得上,而不只是“接口返回了文本”。自动检查可以找出结构问题,交付质量仍需要回听和人工确认。 下文的合成案例是一份12秒录音结果,最后一段到第8秒结束。剩下4秒值得回听,却不能直接算成“漏识别了4秒”:那里也可能只有静音。 检查器没有听过录音。两个说话人标签、或接近终点的时间戳,都不等于转写正确;录音内的匿名标签也不是人物身份。 应用出口示意:已有 FunClip 工作区。不是下文合成样例的运行截图。 先分清四个问题

Changes

No pricing, plan, or positioning changes detected since tracking started.

Design

First-screen visual analysis & :root CSS

Minimal Monochrome (Dark canvas, airy density) featuring split copy & visual layout and interactive hero demo / tool visuals.

First-screen preview · click to zoom

First-screen pitch & conversion flow

“让语音,成为可用的信息。”

Split Copy & Visual hero structure paired with start free / sign up CTA and live generated output showcase.

Conversion path: Split Copy & Visual + Interactive Hero Demo / Tool → Start Free / Sign Up · Trust bar: Live Generated Output Showcase

  • Pain Relief & Replacement

Palette · click to copy

5-D design score7.9 / 10
Typography
8.0
Color & Contrast
8.0
Layout & Spacing
8.0
Craft & Polish
8.0
Distinctiveness
7.0

Structure

Fonts
Inter
Primary visual
Interactive Hero Demo / Tool
Conversion CTA
Start Free / Sign Up
Social proof
Live Generated Output Showcase
Navigation
Full-Width Top Bar
Theme & density
Dark · Airy
Corners & depth
Medium Radius · Hairline borders
Preview full AI design brief (includes :root CSS variables & layout rules)Style tokens only; never copy brand names, logos, copy, or images
# Design Brief: Minimal Monochrome

Monochrome palette with a single accent color, hairline dividers, and generous whitespace.

## Design Tokens

```css
:root {
  --background: #101612;
  --surface: #ffffff;
  --foreground: #f5f7f4;
  --muted-foreground: #8e9c93;
  --accent: #8ed4ae;
  --radius: 8px;
  --font-heading: "Inter", ui-sans-serif, system-ui, sans-serif;
  --font-body: "Inter", ui-sans-serif, system-ui, sans-serif;
}
```

Color area ratio: Background #101612 (~50%) · Surface #FFFFFF (~22%) · Text #F5F7F4 (~10%) · Muted Text #8E9C93 (~10%) · Accent #8ED4AE (~8%). Background and surface dominate the canvas; reserve the accent color for primary buttons and active states.

## Layout & Conversion Structure

- Hero layout: Split Copy & Visual
- Navigation: Full-Width Top Bar
- Theme & density: Dark · Airy
- Border radius & elevation: Medium Radius; no drop shadows—separate layers with hairline borders and whitespace
- Primary visual medium: Interactive Hero Demo / Tool
- Conversion CTA: Start Free / Sign Up · Trust bar uses Live Generated Output Showcase
- Headline hook: Pain Relief & Replacement
- Spatial layout notes: Split Copy & Visual hero structure paired with start free / sign up CTA and live generated output showcase.
- Visual summary: Minimal Monochrome (Dark canvas, airy density) featuring split copy & visual layout and interactive hero demo / tool visuals.

## Avoid

- Multi-stop neon gradients
- Heavy drop shadows
- More than one competing accent color

Implement the UI following the design tokens and layout specifications above: use only these color variables, and reserve the accent color strictly for primary actions and active states. Adopt only this visual style—never copy any reference product’s brand name, logo, copy, or imagery.

Questions & evidence

How much website traffic does FunASR get?
Vaneform estimates 34.6K website visits in 2026-10. Visits can include repeat visits; this estimate measures web traffic separately from paying customers or revenue.Source
What is FunASR and what platforms does it support?
FunASR is catalogued under Transcription & Dictation: 让语音,成为可用的信息。. Supported platforms: CLI / SDK, API.
Is FunASR open source?
FunASR has a public repository (https://github.com/FunAudioLLM/SenseVoice) with its source repository published, recording 2,754 GitHub stars.Source

Related products in this category

Video & Audio / Transcription & Dictation

These products share a catalog category. Compare customer tasks, features and deployment requirements before treating them as alternatives.