Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
bytedance avatar

Byted Las Asr Pro

  • 78 installs
  • 411 repo stars
  • Updated August 4, 2026
  • bytedance/agentkit-samples

byted-las-asr-pro is a Claude skill that transcribes audio and video to text with enhanced analysis via the Volcengine LAS speech recognition service.

About

This skill transcribes audio and video to text using Volcengine LAS enhanced speech recognition (ASR Pro), driven through the las-cli command-line tool. A developer uploads an audio file to TOS storage, and the skill runs recognition with denoising, voice activity detection, punctuation, language and emotion detection, and speaker separation. It estimates the cost first and waits for confirmation before running.

  • Transcribes audio/video to text via Volcengine LAS enhanced ASR
  • Supports denoising, VAD, punctuation, language and emotion detection, speaker diarization
  • Wraps the las-cli operator with a price-estimate confirmation step

Byted Las Asr Pro by the numbers

  • 78 all-time installs (skills.sh)
  • Ranked #824 of 1,335 Generative Media skills by installs in the Skillselion catalog
  • Data as of Aug 5, 2026 (Skillselion catalog sync)
At a glance

byted-las-asr-pro capabilities & compatibility

Requires a Volcengine LAS account (VE_TIP_TOKEN or las.apiKey); billed per-run, estimated via las-cli price before execution

Capabilities
speech to text · transcription · speaker diarization
Works with
openai
Use cases
transcription
Runs
Runs locally
Pricing
Bring your own API key
From the docs

What byted-las-asr-pro says it does

增强语音识别(ASR Pro):将音频/视频转写为文本,支持降噪、说话人分离等增强分析
SKILL.md
las-cli operators call las_asr_pro \
SKILL.md
若 las-cli 命令不存在,则执行 npm i -g "@volcengine/las-cli@0.3.8"
SKILL.md
npx skills add https://github.com/bytedance/agentkit-samples --skill byted-las-asr-pro

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs78
repo stars411
Last updatedAugust 4, 2026
Repositorybytedance/agentkit-samples

What it does

Transcribe audio or video to text with speaker separation and enhanced analysis via Volcengine LAS.

Who is it for?

High-quality speech-to-text with denoising, punctuation, language/emotion detection, and speaker diarization via Volcengine LAS.

When should I use this skill?

A user asks to transcribe a recording or video, generate captions, or separate speakers.

What you get

A transcript with punctuation, language and emotion detection, and speaker separation.

  • text transcript
  • speaker-separated segments

By the numbers

  • las-cli version 0.3.8
  • 4-step workflow plus a Step 0 pre-check
  • mandatory price-estimate confirmation before run

Files

SKILL.mdMarkdownGitHub ↗

ASR Pro 语音识别 (las_asr_pro)

当用户希望对音频进行高质量语音识别(含降噪、VAD、标点、语种识别、情绪识别、说话人分离等)时使用本 Skill。

注意事项(⚠️ 必须遵守)

  • 千万不要让用户输入 AccessKey 和 SecretKey。
  • 下载文件只能使用 las-cli tos download 命令(兼容 tos 和非 tos 资源)。
  • las-cli operators calllas-cli price 命令的输出结果非常重要,必须完整返回给用户。

工作流(严格按步骤执行)

复制此清单并跟踪进度:

执行进度:
- [ ] Step 0: 前置检查
- [ ] Step 1: 初始化与准备
- [ ] Step 2: 预估价格
- [ ] Step 3: 执行任务并呈现结果
- [ ] Step 4: 失败排查

Step 0: 前置检查

版本检查(必须):请执行 las-cli install 0.3.8。若 las-cli 命令不存在,则执行 npm i -g "@volcengine/las-cli@0.3.8"

  • 新版本提醒:请执行 npm info @volcengine/las-cli version 查看当前可用的最新 Skill 版本(las-cli 与 Skill 版本一一对应)。若返回版本高于本 Skill 标注版本,请先提示用户升级当前 Skill。升级方式:请用户发送自然语言指令 帮我安装byted-las-asr-pro技能,等待自动完成安装即可。
  • 认证检查:优先使用运行时自动注入的平台身份;若当前环境未自动注入,再运行 las-cli config show 确认旧版凭证 las.apiKeyvolcengine.region 已配置。

Step 1: 初始化与准备

  • 上传资源: 若输入为本地音频文件,先上传到 TOS。
  # 示例: 上传 input.wav
  las-cli tos upload ./input.wav
  # 假设返回 "tos_url": "tos://my-bucket/inputs/xxx.wav"
  • 准备参数: 创建 params.json,指定输入和输出路径。
  {
    "input_path": "tos://my-bucket/inputs/xxx.wav",
    "output_path": "tos://my-bucket/outputs/asr/"
  }

Step 2: 预估价格(⚠️ 必须获得用户确认)

las-cli price 命令的输出结果非常重要,不要做任何精简、修改或摘要处理。

  • 算价: 必须完整返回以下命令输出的结果(markdown 格式)
  las-cli price las_asr_pro \
    --params-file ./params.json \
    --format markdown
  • 用户确认: 输出预估价格后,必须等待用户确认

Step 3: 执行任务并呈现结果(⚠️ 直接输出命令执行结果,不要精简)

las-cli operators call 命令的输出结果非常重要,不要做任何精简、修改或摘要处理。

  • 执行: 必须完整返回以下命令输出的结果(markdown 格式)
  las-cli operators call las_asr_pro \
    --params-file ./params.json \
    --format markdown \
    --out ./result.md

Step 4: 失败排查

保持原样。

  • 检查任务状态: las-cli task status <task_id> --operator las_asr_pro
  • 检查输入: 确认音频格式、TOS 对象是否存在。
  • 检查输出: 确认 output_path 目录可写。
  • 检查配置: las-cli config show,确认鉴权和 Region。

Related skills

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.